Indoor decoration effect evaluation method and device based on machine vision and geometric algorithm

Abstract

The application provides an indoor decoration effect evaluation method and device based on machine vision and geometric algorithms, wherein the method comprises the following steps: obtaining house characteristic information and furniture characteristic information of a target house, obtaining a space utilization rate score based on furniture area and house area, inputting the furniture characteristic information and a standard furniture library in each space into a furniture richness scoring model, the furniture richness scoring model comprising one or more of a core furniture score, a decorative furniture score and an incorrectly arranged furniture score, and obtaining a furniture richness score by weighted summation of each type of furniture score, inputting the furniture characteristic information and the house characteristic information into a furniture arrangement reasonability scoring model, the furniture arrangement reasonability scoring model comprising one or more of a function area rationality score, a space permeability score and a furniture selection rationality score, and obtaining a furniture arrangement reasonability score by weighted summation of each score, and confirming a decoration effect score of the target house based on the above score items.

Description

Technical Field

[0001] This invention relates to a decoration effect evaluation system and apparatus, and more particularly to an interior decoration effect evaluation method and apparatus based on machine vision and geometric algorithms. Background Technology

[0002] The technical problem to be solved by the present invention is an evaluation scheme for interior decoration effects, which can be used to evaluate the generated decoration effect drawings.

[0003] In the current home decoration market, a large number of designers and intelligent algorithms generate interior design plans for clients every day. The quality of these designs varies greatly. While intelligent algorithms produce a large number of drawings quickly, the design schemes often involve randomness. Therefore, selecting high-quality design schemes from a large pool of options requires significant time and multi-dimensional evaluation. Although there are many solutions on the market that use intelligent algorithms to design schemes (such as CN109740243B), there are no solutions that utilize intelligent algorithms to evaluate the decoration effect. This results in a relatively narrow range of considerations.

[0004] On the other hand, referring to CN109740243B, a method is disclosed that M algorithm models are trained sequentially on M types of furniture. Each model is trained on a single piece of furniture, and an evaluation model is trained for each piece of furniture. The training process must be performed sequentially. After each piece of furniture is arranged, the evaluation model for the next piece of furniture is trained based on the previous model, thus achieving the purpose of scoring the layout position of each piece of furniture and finally obtaining the optimal solution. The advantage of this scheme is that the entire process is fully automated and does not require manual intervention, but the disadvantages are also obvious:

[0005] Disadvantage 1: Model training is complex, requiring M models to be trained for M pieces of furniture. If there are N different types of spaces, then (N1*M1+N2*M2+……+Nn*Mn) models are needed. Disadvantage 2: It requires a large amount of data, has a long training time, and the corresponding GPU training cost is high. Disadvantage 3: The algorithm has poor scalability and maintainability. If a piece of furniture is deleted or added in an intermediate step, the subsequent algorithm models will not be reusable or the effect will drop sharply. It is necessary to recreate the dataset and retrain all subsequent algorithm models to ensure the evaluation quality. Summary of the Invention

[0006] To overcome one or more of the aforementioned drawbacks of previous methods, this invention provides an interior decoration effect evaluation method based on machine vision and geometric algorithms. This method uses machine vision and geometric algorithms to obtain house feature information and furniture information, and ultimately obtains the decoration effect score by calculating space utilization score, furniture richness score, and furniture arrangement rationality score. Compared to known decoration effect evaluation models, this method reduces the complexity of model training, has strong scalability, and improves the accuracy and interpretability of the decoration effect evaluation model.

[0007] This invention provides a method for evaluating the effect of interior decoration based on machine vision and geometric algorithms, comprising:

[0008] Obtain the house feature information and furniture feature information of the target house;

[0009] Space utilization score is obtained based on furniture area and house area;

[0010] The furniture feature information and standard furniture library in each space are input into the furniture richness scoring model. The furniture richness scoring model includes one or more of the following: core furniture score, decorative furniture score, and incorrectly arranged furniture score. The furniture richness score is obtained by weighted summation of the scores of each type of furniture.

[0011] The furniture feature information and house feature information are input into the furniture layout rationality scoring model. The furniture layout rationality scoring model includes one or more of the following: functional area rationality score, spatial permeability score, and furniture selection rationality score. The furniture layout rationality score is obtained by weighted summation of each score.

[0012] Based on the above scoring items, the score for the renovation effect of the target house is determined.

[0013] The method for evaluating interior decoration effects based on machine vision and geometric algorithms provided by the present invention is characterized in that the identification and acquisition of house feature information specifically includes:

[0014] The room element information is obtained based on the instance segmentation model. The wall inflection points are extracted to form parallel wall lines. The common points of the parallel wall lines are connected to form closed loop lines. The wall space polygon is constructed based on the closed loop lines.

[0015] Extract the inflection points of doors and windows to form parallel door and window lines, connect the door and window lines to form door and window polygons, and take the intersection of the door and window lines with the parallel wall lines to obtain the geometric model of the doors and windows;

[0016] Room identification information is obtained based on optical character recognition algorithms. The room identification information includes functional area identification, scale identification, and space area identification.

[0017] The method for evaluating interior decoration effects based on machine vision and geometric algorithms provided by the present invention is characterized by acquiring furniture characteristic information, specifically including:

[0018] Furniture feature bounding boxes are obtained based on object detection algorithms;

[0019] Remove overlapping boundaries from furniture feature bounding boxes based on nonmaximum suppression;

[0020] The furniture feature information and room feature information are associated with and bound to the room identification information.

[0021] The method for evaluating interior decoration effects based on machine vision and geometric algorithms provided by the present invention is characterized in that the acquisition of house feature information and furniture feature information of the target house further includes:

[0022] Geometric models are built based on house and furniture feature information using a general geometric algorithm library;

[0023] A mapping model for the target house feature information and furniture feature information is obtained.

[0024] The method for evaluating interior decoration effects based on machine vision and geometric algorithms provided by the present invention is characterized in that the furniture richness scoring model specifically includes:

[0025]

[0026] Where α represents the scoring weight coefficient of each core furniture piece, β represents the deduction weight coefficient of each missing core furniture piece, γ represents the scoring weight coefficient of each decorative optional furniture piece, θ represents the deduction weight coefficient of incorrectly arranged furniture, F represents each paired core furniture piece (1 if present, 0 otherwise), G represents missing core furniture pieces (-1 if missing, 0 if present), H represents decorative furniture pieces (1 if present, 0 otherwise), and J represents incorrectly arranged furniture pieces (1 if arranged, 0 otherwise).

[0027] The method for evaluating interior decoration effects based on machine vision and geometric algorithms provided by the present invention is characterized in that the rational analysis of functional areas specifically includes:

[0028] Based on the current usage distance of the furniture, determine the rationality of the furniture's usage distance;

[0029] The rationality of the furniture placement is determined by the relative positions of the first and second pieces of furniture.

[0030] The method for evaluating interior decoration effects based on machine vision and geometric algorithms provided by the present invention is characterized in that the spatial transparency analysis specifically includes:

[0031] The activity area is obtained based on the house outline and furniture outline, and the activity area includes the unfolded area of ​​movable parts and the usable area of ​​immovable parts;

[0032] The activity area is defined by the bounding box of the activity area. If the bounding boxes of the activity area overlap, the arrangement is unreasonable. If the bounding boxes of the activity area do not overlap, the arrangement is reasonable.

[0033] The method for evaluating interior decoration effects based on machine vision and geometric algorithms provided by the present invention is characterized in that the rational analysis of furniture selection specifically includes:

[0034] The current furniture space occupancy rate is calculated based on the current furniture area and the current room area. If the occupancy rate is within the preset range, the furniture selection is reasonable; if the occupancy rate is not within the preset range, the furniture selection is unreasonable.

[0035] The present invention also provides an interior decoration effect evaluation device based on machine vision and geometric algorithms, comprising:

[0036] The acquisition unit acquires the house feature information and furniture feature information of the target house;

[0037] The space utilization scoring unit is used to obtain a space utilization score based on furniture area and house area.

[0038] The furniture richness scoring unit is used to input the furniture feature information and standard furniture library in each space into the furniture richness scoring model. The furniture richness scoring model includes one or more of the following: core furniture score, decorative furniture score, and incorrectly arranged furniture score. The furniture richness score is obtained by weighted summation of the scores of each type of furniture.

[0039] The furniture arrangement rationality scoring unit is used to input the furniture feature information and house feature information into the furniture arrangement rationality scoring model. The furniture arrangement rationality scoring model includes one or more of the following: functional area rationality score, spatial permeability score, and furniture selection rationality score. The furniture arrangement rationality score is obtained by weighted summation of the scores.

[0040] The "Confirm Decoration Effect" unit is used to confirm the decoration effect score of the target house based on the above scoring items.

[0041] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the indoor decoration effect evaluation method based on machine vision and geometric algorithms as described above. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0043] Figure 1 This is a flowchart illustrating the decoration effect evaluation method provided by the present invention;

[0044] Figure 2 This is a schematic diagram of the process for obtaining house feature information of a target house provided by the present invention;

[0045] Figure 3 This is a schematic diagram of the process for obtaining furniture feature information of a target house provided by the present invention;

[0046] Figure 4 This is a schematic diagram of the process for obtaining the geometric model of a target house provided by the present invention;

[0047] Figure 5 This is a flowchart illustrating the method for rational analysis of functional areas provided by the present invention;

[0048] Figure 6 This is a flowchart illustrating the spatial permeability analysis method provided by the present invention;

[0049] Figure 7 This is a flowchart illustrating the rationality analysis of furniture selection provided by the present invention;

[0050] Figure 8 This is a schematic diagram of the structure of the interior decoration effect evaluation device based on machine vision and geometric algorithms provided by the present invention;

[0051] Figure 9 This is a block diagram of an electronic device for evaluating the effect of interior decoration based on machine vision and geometric algorithms, as provided by the present invention. Detailed Implementation

[0052] The embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0053] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0054] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0055] When using expressions such as "at least one of A, B, and C," the expression should generally be interpreted in accordance with the meaning commonly understood by a person skilled in the art (e.g., "a system having at least one of A, B, and C" should include, but is not limited to, systems having A alone, having B alone, having C alone, having A and B, having A and C, having B and C, and / or having A, B, and C, etc.). The terms "first" and "second" are used only for naming distinction and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more features.

[0056] Figure 1 This is a flowchart illustrating an interior decoration effect evaluation method based on machine vision and geometric algorithms provided in an embodiment of the present invention. Figure 1 As shown, the method includes:

[0057] Step 110: Obtain the house feature information and furniture feature information of the target house.

[0058] Specifically, the target house can refer to the house to be evaluated. The house to be evaluated can be produced by a designer based on the user's decoration needs and the house type design, or it can be automatically designed by a machine algorithm based on the decoration budget, house area, and user's preferred style.

[0059] The house feature information should at least include the following: the outline coordinates of each room, the coordinates of the doors and windows, and the purpose of each room (kitchen, living room, bedroom, etc.). The furniture feature information should at least include the following: the size and area of ​​the furniture, the type of furniture, the rotation angle, the scaling ratio, and the position coordinates. By extracting this information and building a geometric model, mapping models of the rooms and their wall outlines, doors, windows, and furniture can be generated for evaluating and scoring the house decoration effect.

[0060] Step 120: Obtain a space utilization score based on the furniture area and the house area.

[0061] Specifically, after obtaining the house feature information and furniture feature information of the target house, the house area (based on house feature information) and furniture area (based on furniture feature information) are extracted to calculate the space utilization rate of the target house. The formula for calculating the space utilization rate is as follows:

[0062] Score_usage=α(Area_fur1+Area_fur2+……+Area_furn) / Area_room

[0063] Where Area_fur represents the furniture area in the current room, Area_room represents the total area of ​​the current room, and α represents the weighting coefficient. The method provided in this embodiment of the invention obtains the space utilization rate of the current house. The space utilization rate of the house can well reflect the use of space resources by the decoration effect, while automatically generated decoration plans often fail to make effective use of space resources.

[0064] In one implementation, the furniture area in the current house can be determined using a YOLOv8 object detection model. The furniture is used as the target for object detection, and a set of bounding boxes is selected. In one possible embodiment, non-maximum suppression (NMS) can be used to remove overlapping bounding boxes, thereby filtering out low-confidence detections. The bounding box contains the coordinates (x1, y1, x2, y2) of the top left and bottom right corners, and the furniture area can be calculated using the point coordinates.

[0065] In another possible embodiment, the furniture area in the current house can be obtained by using a deep learning model to extract the furniture information in the room, using it as the input value of the deep learning model, and training the deep learning model with training samples.

[0066] In another possible embodiment, the furniture area in the current house can be obtained by constructing a reinforcement learning model using Actor-Critic. The furniture information is used as the State Space in the reinforcement learning algorithm, the discretized spatial coordinates, rotation angle, and scaling ratio are used as the Action Space in the reinforcement learning, and the output of the environmental feedback model M is used as the Q-value (action value) in the reinforcement learning to obtain the furniture area.

[0067] Step 130: Input the furniture feature information and standard furniture library in each space into the furniture richness scoring model. The furniture richness scoring model includes the core furniture score, decorative furniture score, and incorrectly arranged furniture score. The furniture richness score is obtained by weighted summation of the scores of each type of furniture.

[0068] Specifically, the standard furniture library refers to all the furniture that should be included in the current living space. This furniture includes core furniture and decorative furniture. Core furniture refers to the furniture that must be included in the current living space, while decorative furniture refers to the furniture that can be included. For example, a standard living room core furniture library includes sofas, televisions (projectors), coffee tables, and entryway cabinets, while a standard living room decorative furniture library includes plants and ornaments. Similarly, a standard bedroom core furniture library includes beds, wardrobes, bedside tables, and air conditioners, while a standard living room decorative furniture library includes plants and illustrations.

[0069] In one implementation, the formula for calculating furniture richness is:

[0070]

[0071] Where α represents the scoring weight coefficient of each core furniture piece, β represents the deduction weight coefficient of each missing core furniture piece, γ represents the scoring weight coefficient of each decorative optional furniture piece, θ represents the deduction weight coefficient of incorrectly arranged furniture, F represents each paired core furniture piece (1 if present, 0 otherwise), G represents missing core furniture pieces (-1 if missing, 0 if present), H represents decorative furniture pieces (1 if present, 0 otherwise), and J represents incorrectly arranged furniture pieces (1 if arranged, 0 otherwise).

[0072] In one implementation, the formula for calculating furniture richness also includes incorrectly placed furniture, which refers to furniture information not included in the current space's standard furniture library. For example, the standard bathroom furniture library does not include sofas; if a sofa is included in the bathroom, it is considered incorrectly placed furniture.

[0073] In one implementation, weighting coefficients can be set for different furniture items, with the weighting coefficient of the core furniture library being greater than that of the decorative furniture library.

[0074] In this embodiment, a furniture richness scoring model can be set up for a type of house space. When the furniture in the model changes, such as deleting or adding a piece of furniture, it will not affect the subsequent algorithm model.

[0075] Step 140: Input the furniture feature information and house feature information into the furniture layout rationality scoring model. The furniture layout rationality scoring model includes one or more of the following: functional area rationality score, spatial permeability score, and furniture selection rationality score. The furniture layout rationality score is obtained by weighted summation of each score.

[0076] Specifically, the furniture arrangement rationality scoring model refers to a scoring model set up based on design experience to assess the rationality of furniture arrangement. The furniture arrangement rationality scoring model includes a functional area rationality model, a spatial permeability model, and a furniture selection rationality model. Furthermore, the three scoring models are independent of each other, and the scores do not affect each other. This improves the interpretability of the model and avoids duplicate deductions.

[0077] Through the above embodiments, the furniture arrangement rationality scoring model is based on the designer's experience, making it more objective than manual evaluation by designers and significantly improving the evaluation speed of decoration design schemes. Compared to machine evaluation algorithms, it has stronger scalability and continuous optimization capabilities. Unlike previous algorithms that trained multiple models for a single space, resulting in strong coupling (adding or removing furniture during the design process required retraining subsequent models), this embodiment only requires training one model for the current space, significantly reducing training costs. Changes in furniture information do not affect the training of the remaining models.

[0078] In one implementation, the evaluation function can be manually intervened to adjust different evaluation weights and strategies for different customers, which is more in line with user needs than machine algorithms.

[0079] Step 150: Based on the above scoring items, confirm the decoration effect score of the target house.

[0080] Specifically, the final score for the renovation effect of the target house is obtained based on the scores for space utilization, furniture richness, and furniture arrangement rationality. The renovation effect scoring function is Score = Score_usage + Score_richness + Score_rationality. The layout and renovation effect images with higher scores are obtained by sorting the Score values.

[0081] In one implementation, scoring weight coefficients can be set for different scoring items, with different weight coefficients reflecting the importance of different scoring items. The scoring coefficients can be set according to customer needs.

[0082] Currently, evaluating the decoration effect requires a precise description of the apartment layout. However, the house feature information and furniture feature information in a large number of existing design schemes cannot be directly applied to the current model and cannot be processed uniformly.

[0083] Based on the above embodiments, Figure 2 This is a schematic diagram of the process for obtaining house feature information of a target house provided by the present invention, such as... Figure 2 As shown, it specifically includes:

[0084] Step 210: Obtain room element information based on the instance segmentation model, extract wall inflection points to form parallel wall lines, connect the common points of the parallel wall lines to form closed loop lines, and construct wall space polygons based on the closed loop lines.

[0085] Specifically, the element information of the current room is obtained by segmenting pre-trained YOLOv8-segment instances. Since the wall and door / window structures are relatively simple and have few types, YOLOv8-s is chosen as the initial weights for model training. Training stops when the model converges and the accuracy and recall of each class reach over 90%. In one feasible embodiment, the collected data is divided into training, validation, and test sets in a 7:2:1 ratio. The trained model is tested on the test set data. For classes with low accuracy and recall, the dataset is supplemented, and strategies such as hyperparameter adjustment and model weight adjustment are employed. Training is then repeated on the training set until the accuracy and recall reach the expected levels.

[0086] Room element information refers to the contour points representing the features of a house extracted by the instance segmentation model. Unlike ordinary object detection algorithms, the YOLOv8-segment instance segmentation model can only extract the element point information of the wall frame, but cannot extract the contour information of the wall frame. Typically, the segmentation model extracts M points for the wall contour, takes out four inflection points: (Xmin,Ymin), (Xmin,Ymax), (Xmax,Ymax), (Xmin,Ymin), and forms two pairs of parallel wall lines, Line1 and Line2 (the longer side is the wall line, and the shorter parallel side is the wall thickness), and a closed polygon Polygon = ([X1,Y1],[X2,Y2],[X3,Y3],[X4,Y4],[X1,Y1]), and finally obtains N wall polygons and 2*N wall lines. Connect the 2N wall lines that have a common point to form K sets of closed loops. Construct a spatial polygon Polygon = ([X1,Y1],[X2,Y2],***,[Xn,Yn],[X1,Y1]) for the closed loops to form K simplified spatial outlines.

[0087] Step 220: Extract the inflection points of doors and windows to form parallel door and window lines, connect the door and window lines to form door and window polygons, and take the intersection of the door and window lines with the parallel wall lines to obtain the door and window geometric model;

[0088] Specifically, room element information is obtained based on the instance segmentation model. Typically, the door and window outlines extracted by the segmentation model have M points. Similarly, four inflection points are extracted to construct door and window polygons. The intersection of the door and window polygons and the wall polygons is then taken to obtain the geometric model of the doors and windows on the wall.

[0089] Step 230: Obtain room identification information based on optical character recognition algorithm. The room identification information includes functional area identification, scale identification, and space area identification.

[0090] Specifically, floor plans usually have marking information. In one embodiment, the PaddleOCR engine can be used to recognize digit characters to obtain the digit character area and recognition results.

[0091] Identification information includes: Chinese characters for room names such as guest, dining, hall, bedroom, room, kitchen, room, bathroom, living room, living room, balcony, terrace, etc.; numbers: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9; units: c, m, d; characters: ".", ",", ":", etc.

[0092] In one possible implementation, the trained model is tested on the test set, and strategies such as supplementing the dataset and adjusting parameters are used for categories with low accuracy and recall. The model is then trained again until the accuracy is above 90%.

[0093] In one possible embodiment, unlike obtaining house feature information, furniture feature information inside a house cannot be obtained through an instance segmentation model. This invention provides a state-of-the-art object detection algorithm to obtain furniture features, while solving the problem of defining overlapping boundaries in the object detection algorithm.

[0094] Based on the above embodiments, Figure 3 This is a schematic diagram of the process for obtaining furniture feature information of a target house provided by the present invention, such as... Figure 3 As shown, the step of obtaining furniture feature information of the target house includes:

[0095] Step 310: Obtain furniture feature bounding boxes based on object detection algorithms.

[0096] Specifically, furniture features include (furniture type, quantity, size, rotation angle, scaling ratio, position coordinates, etc.). The furniture feature bounding box refers to the position and size of the furniture defined by the object detection algorithm. This feature bounding box generally includes: top-left corner coordinates (x1, y1): representing the position of the top-left vertex of the rectangle in the image or space; bottom-right corner coordinates (x2, y2): representing the position of the bottom-right vertex of the rectangle in the image or space; width: equal to x2 - x1, representing the width of the rectangle; and height: equal to y2 - y1, representing the height of the rectangle.

[0097] In one implementation, the YOLOv8 object detection algorithm can be used to obtain the furniture feature bounding boxes. If there are many types of furniture, the weight model can choose YOLOv8-x to obtain the furniture feature bounding boxes. No limitation is made here.

[0098] Step 320: Remove overlapping boundaries in the furniture feature bounding box based on nonmaximum suppression.

[0099] Specifically, nonmaximum suppression is a technique used to remove overlapping bounding boxes, ensuring that each object is detected only once.

[0100] In one implementation, the detection results can be sorted in descending order of confidence score. The detection result with the highest score is selected as the baseline and retained. The remaining detection results are then iterated through, and any box with an IoU (Intersection over Union) greater than a certain threshold (e.g., 0.5) with the baseline box is discarded. The next highest-scoring unprocessed detection result is then selected.

[0101] In one implementation, overlapping boundaries between the core furniture and decorative furniture bounding boxes are removed based on nonmaximum suppression. Specifically, if the core furniture features overlap with the decorative furniture bounding box, for example, if a decorative furniture vase is placed on the core furniture coffee table, causing the detected furniture bounding boxes to overlap, the detection result with the highest score is selected and the detection result with the reference box IoU (Intersection over Union) greater than 0.5 is discarded.

[0102] Step 330: Assign and bind the furniture feature information and room feature information to the room identification information.

[0103] Specifically, the room name information, area information, and scale identified within the spatial outline are associated with the space. The scale of the room identification is converted proportionally to obtain the actual dimensions of each wall, door, window, and furniture. If the floor plan does not have scale information, the scale information is calculated by combining the outline area with the identified actual area to obtain the actual dimensions. Through the above embodiments, the actual dimensions of the house features and furniture features can be obtained.

[0104] Based on the above embodiments, the house features and furniture features of the current house can be obtained, but they cannot be directly used to evaluate the decoration effect. This embodiment provides a method that uses a general geometric algorithm library, such as OCC, CGAL, shapely, etc., to build a model of the current house features.

[0105] Based on the above examples Figure 4 This is a schematic diagram of the process for obtaining the geometric model of a target house provided by the present invention, as shown below. Figure 4 As shown, it specifically includes:

[0106] Step 410: Establish a geometric model based on the house feature information and furniture feature information using a general geometric algorithm library.

[0107] Specifically, the general-purpose geometric algorithm library refers to software tools used to process and manipulate geometric data. Some common general-purpose geometric algorithm libraries include OCC, CGAL, and shapely. A geometric model is established based on the positional information of each element point using this general-purpose geometric algorithm library.

[0108] Step 420: Obtain the mapping relationship model between the target house feature information and the furniture feature information.

[0109] Specifically, the mapping model describes the relationships between various elements within a room, including the room's geometry, the relative positions and interactions between walls, doors, windows, and furniture. The mapping relationships are output as a JSON file and can be embedded as a submodule into a decoration effect evaluation system to assist home decoration design and improve work efficiency.

[0110] Based on the above embodiments, Figure 5 This is a flowchart illustrating the method for rational analysis of functional areas provided by the present invention, as shown below. Figure 5 As shown, it specifically includes:

[0111] Step 510: Determine the reasonableness of the furniture's usage distance based on the current usage distance.

[0112] Specifically, the usage distance of the furniture refers to the distance that needs to be walked during use. For example, if the dining table needs to be placed next to the kitchen, then take the dining table polygon and the kitchen polygon, and calculate whether the distance between the two polygons is less than a threshold. Similarly, calculate the distance between the sideboard and the dining table, the distance between the entryway cabinet and the entrance door, and calculate the positions of the coffee table geometric model and the sofa geometric model to determine whether it is reasonable.

[0113] Step 520: Based on the relative positions of the first and second pieces of furniture, determine the rationality of the furniture placement.

[0114] Specifically, the relative placement of furniture refers to the relationship between the placement of two related pieces of furniture. For example, based on experience, a sofa needs to face the TV for viewing. The intersection of the sofa polygon and the TV polygon with the wall line can be used to obtain the back lines of the sofa and TV. The parallelism of the two back lines can be calculated to determine whether the sofa and TV arrangement is reasonable.

[0115] In one implementation, a functional area rational analysis model can be established for each space based on design experience and user needs. Score1 = k * α, where k = 1 for a reasonable layout and k = -1 for an unreasonable layout, and α is the weight score.

[0116] In one possible embodiment, commercially available machine learning algorithms cannot calculate the openness of home spaces, resulting in situations where, although furniture is arranged reasonably, the spatial distance between furniture is too narrow, hindering normal furniture use. For example, the geometric model of furniture may obstruct the opening of doors, windows, and cabinets, and there may not be sufficient aisle clearance between furniture pieces. Furthermore, collision detection between furniture and the house can only detect overlap, not activity space. Based on this, this embodiment provides a method for analyzing spatial openness.

[0117] Based on the above embodiments, Figure 6 This is a flowchart illustrating the spatial permeability analysis method provided by the present invention, as shown below. Figure 6 As shown, it specifically includes:

[0118] Step 610: Obtain the activity area based on the house outline and furniture outline, wherein the activity area includes the unfolded area of ​​movable parts and the usable area of ​​immovable parts.

[0119] Specifically, the house outline and furniture outline are obtained through step 110. The movable area refers to the unfolded area of ​​the movable parts in the current house and the usable area of ​​the current furniture. The calculation of the unfolded area and the usable area of ​​the furniture can be preset according to the characteristics of the house and the furniture. The calculation of the unfolded area of ​​the movable parts specifically includes identifying the door and window position information of the current house through step 110, and using the door and window width as the length of the unfolded area, A = (1 / 2)*r^2*θ, where A is the area of ​​the arc, r is the door and window width, and θ is the radius in radians.

[0120] The calculation of the current furniture's usable area specifically includes identifying the furniture's outline information through step 110, and increasing the furniture's length and width accordingly based on the identified furniture type as the length and width of the current furniture's usable area.

[0121] Step 620: Define the bounding box of the activity area based on the activity area. If there is an intersection between the bounding boxes of the activity area, the arrangement is unreasonable. If there is no intersection between the bounding boxes of the activity area, the arrangement is reasonable.

[0122] Specifically, the bounding box is a method for finding the optimal solution for spatial collisions. The bounding box algorithm can be an AABB (Axis-aligned bounding box), a sphere, or a line-segment / line-segment or line-segment / polygon collision detection; no specific limitation is made here. The active area bounding box is defined by the active area, and then it is determined whether the AABBs of two pieces of furniture intersect. If there is an intersection, then the two objects are considered to have collided. The active area bounding box represents the minimum and maximum values ​​of each coordinate axis. For floor plans, the Z-axis value is uniformly set to 0. If there is an intersection between active area bounding boxes, K = -1; if there is no intersection, K = 1. The spatial permeability score Score2 = k * β, where β is the weight score.

[0123] Currently in the home decoration design field, the selection of furniture size is mainly based on the designer's design experience to choose matching furniture size. There is a lack of evaluation standards for furniture size that is too large or too small. For example, a large space may be equipped with small furniture; a small space may be equipped with a large sofa or bed, making it impossible to place other furniture such as wardrobes and televisions.

[0124] Based on the above embodiments, Figure 7 This is a flowchart illustrating the rationality analysis of furniture selection provided by the present invention, as shown below. Figure 7 As shown, it specifically includes:

[0125] Step 710: Calculate the current furniture space occupancy rate based on the current furniture area and the current room area.

[0126] Specifically, the current furniture area and the current room area can be calculated in step 110, wherein the formula for the current furniture space occupancy rate is:

[0127] Score_usage=α(Area_fur / Area_room)

[0128] Where Area_fur represents the furniture area in the current room, Area_room represents the current room area, and α represents the weighting coefficient. The method provided in this embodiment of the invention obtains the proportion of room area occupied by the current furniture. This proportion can effectively reflect the rationality of furniture selection. It should be noted that the calculation of the current furniture space occupancy rate generally focuses on the area of ​​core furniture; the area of ​​decorative furniture is generally not calculated.

[0129] Step 720: If the occupancy rate is within the preset range, the furniture selection is reasonable; if the occupancy rate is not within the preset range, the furniture selection is unreasonable.

[0130] Specifically, when the furniture occupancy rate is within a preset threshold, it indicates that the furniture selection is appropriate and reasonable. When the furniture occupancy rate exceeds the preset threshold, the furniture selection is too large and unreasonable. When the furniture occupancy rate is below the preset threshold, the furniture selection is too small and unreasonable. The preset threshold range is set in advance.

[0131] Based on any of the above embodiments, the present invention provides an interior decoration effect evaluation device based on machine vision and geometric algorithms. Figure 8 This is a schematic diagram of the interior decoration effect evaluation device based on machine vision and geometric algorithms provided by the present invention, as shown below. Figure 8 As shown, the device includes:

[0132] Unit 810 acquires the house feature information and furniture feature information of the target house;

[0133] Space utilization scoring calculation unit 820 is used to obtain a space utilization score based on furniture area and house area;

[0134] The furniture richness scoring unit 830 is used to input the furniture feature information and standard furniture library in each space into the furniture richness scoring model. The furniture richness scoring model includes one or more of the following: core furniture score, decorative furniture score, and incorrectly arranged furniture score. The furniture richness score is obtained by weighted summation of the scores of each type of furniture.

[0135] The furniture arrangement rationality scoring unit 840 is used to input the furniture feature information and house feature information into the furniture arrangement rationality scoring model. The furniture arrangement rationality scoring model includes one or more of the following: functional area rationality score, spatial permeability score, and furniture selection rationality score. The furniture arrangement rationality score is obtained by weighted summation of the scores.

[0136] The decoration effect confirmation unit 850 is used to confirm the decoration effect score of the target house based on the above scoring items.

[0137] The device provided in this invention calculates the house feature information and furniture feature information of the target house respectively, and then calculates the space utilization rate score, furniture richness score, and furniture arrangement rationality score based on the above feature information, finally obtaining the decoration effect score of the target house. Compared with known decoration effect evaluation models, this reduces the complexity and cost of model training. On the other hand, this device has strong scalability and maintainability.

[0138] Based on any of the above embodiments, the acquisition unit 810 is specifically used for:

[0139] The system identifies and obtains house feature information, obtains room element information based on instance segmentation model, extracts wall inflection points to form parallel wall lines, connects the common points of parallel wall lines to form closed loop lines, and constructs wall space polygons based on closed loop lines.

[0140] Extract the inflection points of doors and windows to form parallel door and window lines, connect the door and window lines to form door and window polygons, and take the intersection of the door and window lines with the parallel wall lines to obtain the geometric model of the doors and windows;

[0141] Room identification information is obtained based on optical character recognition algorithms. The room identification information includes functional area identification, scale identification, and space area identification.

[0142] Obtain furniture feature information and obtain furniture feature bounding boxes based on object detection algorithms;

[0143] Remove overlapping boundaries from furniture feature bounding boxes based on nonmaximum suppression;

[0144] The furniture feature information and room feature information are associated with and bound to the room identification information.

[0145] Construct a mapping relationship model, and establish a geometric model based on the house feature information and furniture feature information using a general geometric algorithm library;

[0146] A mapping model for the target house feature information and furniture feature information is obtained.

[0147] Based on any of the above embodiments, the furniture richness scoring unit 830 is specifically used for:

[0148] The scores for core furniture, decorative furniture, and misplaced furniture are calculated to obtain the final furniture richness score.

[0149] Based on any of the above embodiments, the furniture arrangement rationality scoring unit 840 is specifically used for:

[0150] Obtain a functional area rationality score, and determine the rationality of the furniture's usage distance based on the current usage distance of the furniture;

[0151] The rationality of the furniture placement is determined by the relative positions of the first and second pieces of furniture.

[0152] A spatial transparency analysis is performed to obtain the activity area based on the building outline and furniture outline. The activity area includes the unfolded area of ​​movable parts and the usable area of ​​immovable parts.

[0153] The activity area is defined by the bounding box of the activity area. If the bounding boxes of the activity area overlap, the arrangement is unreasonable. If the bounding boxes of the activity area do not overlap, the arrangement is reasonable.

[0154] The furniture selection analysis is based on the current furniture area and the current room area to calculate the current furniture space occupancy rate. If the occupancy rate is within the preset range, the furniture selection is reasonable; if the occupancy rate is not within the preset range, the furniture selection is unreasonable.

[0155] Figure 9 A block diagram schematically illustrates an electronic device 1000 suitable for implementing an interior decoration effect evaluation method based on machine vision and geometric algorithms according to embodiments of the present disclosure.

[0156] like Figure 9 As shown, an electronic device 1000 according to an embodiment of the present disclosure includes a processor 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage portion 1008 into a random access memory (RAM) 1003. The processor 1001 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 1001 may also include onboard memory for caching purposes. The processor 1001 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0157] RAM 1003 stores various programs and data required for the operation of electronic device 1000. Processor 1001, ROM 1002, and RAM 1003 are interconnected via bus 1004. Processor 1001 performs various operations of the method flow according to embodiments of the present disclosure by executing programs in ROM 1002 and / or RAM 1003. It should be noted that the programs may also be stored in one or more memories other than ROM 1002 and RAM 1003. Processor 1001 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in said one or more memories.

[0158] According to embodiments of this disclosure, the electronic device 1000 may further include an input / output (I / O) interface 1005, which is also connected to a bus 1004. The electronic device 1000 may also include one or more of the following components connected to the I / O interface 1005: an input section 1006 including a keyboard, mouse, etc.; an output section 1007 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN card, modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as needed. A removable medium 1011, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 1010 as needed so that computer programs read from it can be installed into the storage section 1008 as needed.

[0159] This disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs that, when executed, implement the method according to the embodiments of this disclosure.

[0160] According to embodiments of this disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, such as including, but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this disclosure, the computer-readable storage medium may include ROM 1002 and / or RAM 1003 and / or one or more memories other than ROM 1002 and RAM 1003 described above.

[0161] Embodiments of this disclosure also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to cause the computer system to implement the methods provided in the embodiments of this disclosure.

[0162] When the computer program is executed by the processor 1001, it performs the functions defined in the system / apparatus of this disclosure embodiments. According to embodiments of this disclosure, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0163] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and may be downloaded and installed via the communication section 1009, and / or installed from a removable medium 1011. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0164] In such an embodiment, the computer program can be downloaded and installed from a network via communication section 1009, and / or installed from removable medium 1011. When the computer program is executed by processor 1001, it performs the functions defined in the system of this disclosure embodiment. According to embodiments of this disclosure, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0165] According to embodiments of this disclosure, program code for executing the computer programs provided in embodiments of this disclosure can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can execute entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0166] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0167] Those skilled in the art will understand that the features described in the various embodiments and / or claims of this disclosure can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this disclosure. In particular, the features described in the various embodiments and / or claims of this disclosure can be combined and / or combined in various ways without departing from the spirit and teachings of this disclosure. All such combinations and / or combinations fall within the scope of this disclosure.

[0168] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. The scope of this disclosure is defined by the appended claims and their equivalents. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of this disclosure, and all such substitutions and modifications should fall within the scope of this disclosure.

Claims

1. A method for evaluating the effect of interior decoration based on machine vision and geometric algorithms, characterized in that: Obtain the house feature information and furniture feature information of the target house; Identifying and acquiring housing feature information, specifically including: The room element information is obtained based on the instance segmentation model. The wall inflection points are extracted to form parallel wall lines. The common points of the parallel wall lines are connected to form closed loop lines. The wall space polygon is constructed based on the closed loop lines. Extract the inflection points of doors and windows to form parallel door and window lines, connect the door and window lines to form door and window polygons, and take the intersection of the door and window lines with the parallel wall lines to obtain the geometric model of the doors and windows; Room identification information is obtained based on optical character recognition algorithm. The room identification information includes functional area identification, scale identification and space area identification. Space utilization score is obtained based on furniture area and house area; The furniture feature information and standard furniture library in each space are input into the furniture richness scoring model. The furniture richness scoring model includes one or more of the following: core furniture score, decorative furniture score, and incorrectly arranged furniture score. The furniture richness score is obtained by weighted summation of the scores of each type of furniture. The furniture feature information and house feature information are input into the furniture layout rationality scoring model. The furniture layout rationality scoring model includes one or more of the following: functional area rationality score, spatial permeability score, and furniture selection rationality score. The furniture layout rationality score is obtained by weighted summation of each score. Based on the above scoring items, the score for the renovation effect of the target house is determined.

2. The method for evaluating interior decoration effects based on machine vision and geometric algorithms according to claim 1, characterized in that, Obtain furniture characteristic information, specifically including: Furniture feature bounding boxes are obtained based on object detection algorithms; Remove overlapping boundaries from furniture feature bounding boxes based on nonmaximum suppression; The furniture feature information and room feature information are associated with and bound to the room identification information.

3. The method for evaluating interior decoration effects based on machine vision and geometric algorithms according to claim 2, characterized in that... The acquisition of the target house's house feature information and furniture feature information also includes: Geometric models are built based on house and furniture feature information using a general geometric algorithm library; A mapping model for the target house feature information and furniture feature information is obtained.

4. The method for evaluating interior decoration effects based on machine vision and geometric algorithms according to claim 1, characterized in that... The furniture richness scoring model specifically includes: Where α represents the scoring weight coefficient of each core furniture piece, β represents the deduction weight coefficient of each missing core furniture piece, γ represents the scoring weight coefficient of each decorative optional furniture piece, θ represents the deduction weight coefficient of incorrectly arranged furniture, F represents each paired core furniture piece (1 if present, 0 otherwise), G represents missing core furniture pieces (-1 if missing, 0 if present), H represents decorative furniture pieces (1 if present, 0 otherwise), and J represents incorrectly arranged furniture pieces (1 if arranged, 0 otherwise).

5. The method for evaluating interior decoration effects based on machine vision and geometric algorithms according to claim 1, characterized in that, The rational analysis of the functional areas specifically includes: Based on the current usage distance of the furniture, determine the rationality of the furniture's usage distance; The rationality of the furniture placement is determined by the relative positions of the first and second pieces of furniture.

6. The method for evaluating interior decoration effects based on machine vision and geometric algorithms according to claim 1, characterized in that, The spatial permeability analysis specifically includes: The activity area is obtained based on the house outline and furniture outline, and the activity area includes the unfolded area of ​​movable parts and the usable area of ​​immovable parts; The activity area is defined by the bounding box of the activity area. If the bounding boxes of the activity area overlap, the arrangement is unreasonable. If the bounding boxes of the activity area do not overlap, the arrangement is reasonable.

7. The method for evaluating interior decoration effects based on machine vision and geometric algorithms according to claim 1, characterized in that, The analysis of reasonable furniture selection specifically includes: The current furniture space occupancy rate is calculated based on the current furniture area and the current room area. If the occupancy rate is within the preset range, the furniture selection is reasonable; if the occupancy rate is not within the preset range, the furniture selection is unreasonable.

8. An interior decoration effect evaluation device based on machine vision and geometric algorithms, characterized in that, include: The acquisition unit acquires the house feature information and furniture feature information of the target house; Identifying and acquiring housing feature information, specifically including: Room element information is obtained based on the instance segmentation model, wall inflection points are extracted to form parallel wall lines, common points of parallel wall lines are connected to form closed loop lines, and wall space polygons are constructed based on the closed loop lines. Extract the inflection points of doors and windows to form parallel door and window lines, connect the door and window lines to form door and window polygons, and take the intersection of the door and window lines with the parallel wall lines to obtain the geometric model of the doors and windows; Room identification information is obtained based on optical character recognition algorithm. The room identification information includes functional area identification, scale identification and space area identification. The space utilization scoring unit is used to obtain a space utilization score based on furniture area and house area. The furniture richness scoring unit is used to input the furniture feature information and standard furniture library in each space into the furniture richness scoring model. The furniture richness scoring model includes one or more of the following: core furniture score, decorative furniture score, and incorrectly arranged furniture score. The furniture richness score is obtained by weighted summation of the scores of each type of furniture. The furniture arrangement rationality scoring unit is used to input the furniture feature information and house feature information into the furniture arrangement rationality scoring model. The furniture arrangement rationality scoring model includes one or more of the following: functional area rationality score, spatial permeability score, and furniture selection rationality score. The furniture arrangement rationality score is obtained by weighted summation of the scores. The "Confirm Decoration Effect" unit is used to confirm the decoration effect score of the target house based on the above scoring items.

9. A computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the method of any one of claims 1 to 7.

Images