Video recording system, video access method and computer readable medium

Abstract

The invention discloses a video recording system, an image access method and a computer readable medium. The video recording system comprises a camera module, a memory device and a processing unit. The memory device stores at least one instruction code. The processing unit is electrically connected with the camera module and the memory device. When the processing unit reads at least one instruction code, at least the following steps are executed. A plurality of images are captured by the camera module to generate a film. At least one universal object in a plurality of frames in a video is identified. An image scene is determined according to the at least one universal object. Performing specific object detection on the frames according to the image scene to identify whether at least one specific object or at least one specific event appears in the frames; a label is attached to at least one of the frames having at least one specific object or at least one specific event. The user screens the image segments covering the specific objects in the specific image scene according to the labels, so that the difficulty of searching the specific objects / events in a large number of accumulated recorded films by the user is reduced.

Description

Technical Field

[0001] This case relates to a video recording system, an image storage method, and a computer-readable medium, and more particularly to a video recording system, an image storage method, and a computer-readable medium that can be used to record videos for object detection. Background Art

[0002] Modern life is filled with a variety of video recording devices, such as cameras, dashcams, surveillance cameras, and body cameras. However, the biggest problem with these products is that they require manual control to start and stop recording. Even the subsequent recorded videos require significant time to organize, using advanced image processing software to slowly review each segment and identify meaningful clips for editing. This often leads to users losing interest, leaving the purchased products unused, or saving the recorded videos to hard drives in the hope of needing them later, but ultimately abandoning them due to the inability to store the data.

[0003] Therefore, how to provide a video recording system to solve the above problems is an important issue in this field. Summary of the Invention

[0004] The present disclosure aims to provide a video recording system, an image access method, and a computer-readable medium, so as to enable a user to filter image segments covering specific objects in a specific image scene according to a plurality of tags.

[0005] The present disclosure provides a video recording system comprising a camera module, a memory device, and a processing unit. The memory device stores at least one instruction code. The processing unit is electrically connected to the camera module and the memory device. When the processing unit reads at least one instruction code, at least the following steps are performed: a plurality of images are captured by the camera module to generate a video. At least one common object is identified in a plurality of frames in the video. An image scene is determined based on the at least one common object. Based on the image scene, a specific object detection is performed on the plurality of frames to identify whether at least one specific object or at least one specific event appears in the plurality of frames. A label is attached to at least one of the plurality of frames having at least one specific object or at least one specific event.

[0006] In one embodiment, the video recording system further includes: a storage device electrically connected to the processing unit, and storing a first neural network and a plurality of second neural networks, wherein the processing unit further performs the following steps: using the first neural network to perform general object detection on the plurality of frames in the video to identify at least one general object in the plurality of frames; determining the image scene based on the at least one general object; selecting one of the plurality of second neural networks based on the image scene; and using the one of the plurality of second neural networks to perform specific object detection on the plurality of frames to identify at least one specific object or specific event in the plurality of frames.

[0007] In one embodiment, the video recording system further includes a motion sensor electrically connected to the processing unit, and sensing the surrounding environment to generate a motion sensing signal. When the processing unit receives the motion sensing signal, it controls the camera module to start up to generate a video.

[0008] In one embodiment, the video recording system further includes an acceleration sensor electrically connected to the processing unit and sensing acceleration information of the camera module, wherein the processing unit controls the camera module to stop filming according to the acceleration information and the motion sensing signal.

[0009] In one embodiment, the video recording system further includes: a network module electrically connected to the processing unit and transmitting data to a cloud server, wherein the processing unit transmits at least one frame with at least one tag to the cloud server based on the acceleration information and the motion sensing signal.

[0010] This disclosure provides an image access method comprising the following steps: performing universal object detection on multiple frames in a video to identify at least one universal object in the frames; determining an image scene based on the at least one universal object; performing specific object detection on the frames based on the image scene to identify whether at least one specific object or at least one specific event is present in the frames; and attaching at least one label to the frames containing the at least one specific object or at least one specific event when at least one specific object is detected in the frames. The labeled frames are then stored.

[0011] In one embodiment, the image access method further includes: storing the frame with at least one tag on a cloud server.

[0012] In one embodiment, the image access method further includes: determining an image scene based on a category of at least one common object and an area ratio of a bounding box, wherein the image scene is one of a road scene, a shopping scene, a travel scene, and a work scene.

[0013] In one embodiment, the image access method further includes: when the number of frames in which at least one specific object appears is 1, storing the frames as an image file; and when the number of frames in which at least one specific object appears is greater than 1, editing the video into a dynamic image file or a short video file so that it contains frames with at least one specific object.

[0014] In one embodiment, the image access method further includes: displaying all stored tags to the user through a user interface; when the user clicks at least one tag, one of the image file, dynamic image file and short video file corresponding to the at least one tag is displayed, wherein the image file, dynamic image file and short video file include frames with at least one specific object or at least one specific event.

[0015] In one embodiment, the at least one specific object is one of a traffic sign, a natural landmark, an artificial landmark, a product trademark, a product item, a store name, and a vehicle.

[0016] In one embodiment, the at least one specific event includes one of the own traffic behavior, surrounding traffic behavior, and sports behavior.

[0017] This disclosure provides a computer-readable medium storing at least one instruction code. When the at least one instruction code is read, the computer-readable medium performs at least the following steps: Performing universal object detection on multiple frames in a video to identify at least one universal object in the multiple frames. Determining the image scene based on the at least one universal object. Performing specific object detection on the multiple frames based on the image scene to identify whether at least one specific object or at least one specific event is present in the multiple frames. When at least one specific object is detected in the multiple frames, at least one label is attached to the frame containing the at least one specific object or at least one specific event. Storing the labeled frames.

[0018] In one embodiment, when the at least one instruction code is read by the processing unit, the processing unit is further caused to execute the following steps: connecting to a cloud server; and storing the frame with the at least one tag in the cloud server.

[0019] In summary, the video recording system disclosed herein can automatically analyze specific objects or events in a video and attach tags to the video frames. This allows users to filter video clips containing specific objects within a specific scene based on these tags, thereby reducing the difficulty of finding specific objects / events within a large number of recorded videos, thereby increasing the value of recorded videos and simplifying the video editing process. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] To make the above and other objects, features, advantages and embodiments of the present disclosure more apparent and understandable, the accompanying drawings are described as follows:

[0021] Figure 1A is a schematic diagram of a video recording system according to an embodiment of the present disclosure;

[0022] Figure 1B is a schematic diagram of a video recording system according to an embodiment of the present disclosure;

[0023] Figure 2 is a flow chart of an image access method of a video recording system according to an embodiment of the present disclosure;

[0024] Figure 3A as well as Figure 3B is a flow chart of some steps of an image access method according to an embodiment of the present disclosure;

[0025] Figures 4A to 4C is a schematic diagram of a road scene according to an embodiment of the present disclosure;

[0026] Figures 5A to 5C is a schematic diagram of a shopping scene according to an embodiment of the present disclosure;

[0027] Figures 6A to 6B FIG. 1 is a schematic diagram of a landscape scene according to an embodiment of the present disclosure.

[0028]

Explanation of symbols

[0029] To make the above and other objects, features, advantages and embodiments of the present disclosure more apparent and understandable, the accompanying symbols are explained as follows:

[0030] 100: Video recording system

[0031] 110: Processing unit

[0032] 112: Memory Device

[0033] 114: Storage device

[0034] 120: Camera module

[0035] 122: Microphone Array

[0036] 130: Network module

[0037] 140:Sensor Hub

[0038] 142: Microcontroller

[0039] 144: Gravity sensor

[0040] 146: Acceleration sensor

[0041] 148: Motion sensor

[0042] 152: Charging port

[0043] 154:Charge Controller

[0044] 156:Battery

[0045] 200: Image access method

[0046] COM, SPE1, SPE2, SPE3, SPE4: Neural Networks

[0047] CS: Cloud Server

[0048] AP: Access Point

[0049] S210~S250, S231~S234, S241~S244: Steps DETAILED DESCRIPTION

[0050] The following examples are described in detail with accompanying illustrations. However, the examples provided are not intended to limit the scope of this disclosure, and the description of the structural operation is not intended to limit the order of execution. Any device with equivalent functionality resulting from the reconfiguration of the components is within the scope of this disclosure. Furthermore, the illustrations are for illustrative purposes only and are not drawn to scale. To facilitate understanding, identical or similar components will be designated with the same reference numerals throughout the following description.

[0051] Unless otherwise noted, terms used throughout the specification and claims generally have their ordinary meanings in the art, within the context of this disclosure, and within the specific context. Furthermore, the terms "comprising," "including," "having," and "containing," etc., as used herein, are open-ended terms meaning "including, but not limited to," and "and / or" as used herein encompass any and all combinations of one or more of the listed items.

[0052] See also Figure 1A , Figure 1A FIG. 1 is a schematic diagram of a video recording system 100 according to an embodiment of the present disclosure. Figure 1AAs shown, the recording system 100 includes a processing unit 110, a camera module 120, a network module 130, and a sensor hub 140. In some embodiments, the processing unit 110 may be implemented as an integrated circuit consisting of a central processing unit (CPU) and a graphics processing unit (GPU). In some embodiments, the processing unit 110 may be considered as the CPU and GPU in a system-on-chip (SoC) with a graphics processing unit (GPU). In some embodiments, the processing unit 110 is electrically connected to the camera module 120, the sensor hub 140, and the network module 130. In some embodiments, the camera module 120 may be implemented as a camera module. In some embodiments, the camera module 120 may include a camera with a high frame rate (e.g., a frame rate of 120 Hz), a high resolution (e.g., a resolution of 4 megapixels (1920x1080) or 12 megapixels (4032x3040)), and a wide viewing angle (e.g., a 120-degree field of view). In some embodiments, the camera specifications of the camera module 120 may be selected based on the amount of data to be recorded and temporarily stored, but the present invention is not limited thereto. In some embodiments, the sensor hub 140 is capable of acquiring, controlling, processing, and integrating sensing information from one or more sensors and providing the processed sensing information to the processing unit 110. In some embodiments, the network module 130 provides mobile network connectivity and GPS location information and connects to the cloud server CS via the internet. In some embodiments, the network module 130 includes a network control unit (CCU), a wireless communication interface, and / or a GPS receiver. In some embodiments, the network module 130 has mobile network connectivity and can provide GPS location information. In some embodiments, the network module 130 transmits data to the cloud server CS via a network. In some embodiments, the network control unit controls the wireless communication interface to connect to an access point (AP), thereby connecting to the cloud server CS via the AP. The image and picture data is then transmitted to the cloud server CS via the network for storage, allowing users to view and manage the data through mobile phone applications and computer programs.

[0053] See also Figure 1B . Figure 1B FIG. 1 is a schematic diagram of a video recording system 100 according to an embodiment of the present disclosure. In some embodiments, the video recording system 100 further includes a memory device 112 , a storage device 114 , a microphone array 122 , a battery 156 , a charging controller 154 , and a charging interface 152 .

[0054] In some embodiments, the memory device 112 and / or the storage device 114 may be implemented as an electrical, magnetic, or optical memory device, or other storage device that stores instructions or data. In some embodiments, the memory device 112 and / or the storage device 114 may be implemented as a volatile memory or a non-volatile memory. In some embodiments, the memory device 112 and / or the storage device 114 may be implemented as a random access memory, a dynamic random access memory, a magnetoresistive random access memory, a phase change memory, or other storage device.

[0055] In some embodiments, the sensor hub 140 includes a microcontroller 142. Furthermore, the sensor hub 140 may also include at least one sensor, such as at least one of an accelerometer 146, a motion sensor 148, and a gravity sensor 144. In this embodiment, the gravity sensor 144, the accelerometer 146, and the motion sensor 148 are electrically connected to the microcontroller 142. In some embodiments, the gravity sensor 144 and the accelerometer 146 are used to sense the motion of the recording system 100 itself, while the motion sensor 148 is used to sense the motion of the surrounding environment. In some embodiments, the microcontroller 142 processes the sensing results of the gravity sensor 144, the accelerometer 146, and the motion sensor 148 and transmits them to the processing unit 110. The processing unit 110 then controls the start and end times of recording for the camera module 120 based on the sensing results.

[0056] In some embodiments, the motion sensor 148 is electrically connected to the processing unit 110 via the microcontroller 142 and senses the motion of people or objects in the surrounding environment to generate a motion sensing signal. In some embodiments, the motion sensor 148 may be implemented as an infrared sensor. In some embodiments, the motion sensor 148 is used to detect whether a person or object is moving nearby to generate a motion sensing signal. That is, the motion sensing signal includes information about whether a person or object is moving near the environment where the recording system 100 is located. In some embodiments, if the processing unit 110 determines that a person or object is moving near the environment where the recording system 100 is located based on the motion sensing signal generated by the motion sensor 148, the processing unit 110 controls the camera module 120 to activate. In some embodiments, when the processing unit 110 receives the motion sensing signal generated by the motion sensor 148, the processing unit 110 controls the camera module 120 to activate and generate / record a video.

[0057] In some embodiments, the acceleration sensor 146 is electrically connected to the processing unit 110 via the microcontroller 142 and senses acceleration information from the camera module 120. In some embodiments, when the camera module 120 is recording images, the processing unit 110 controls the acceleration sensor 146 to sense the acceleration information from the camera module 120. In some embodiments, the processing unit 110 can determine the motion state of the recording system 100 (for example, if the recording system 100 is installed in a moving vehicle and the acceleration information generated by the acceleration sensor 146 includes fluctuating acceleration values) and / or the motion state of people or objects near the recording system 100 based on the motion sensing signal generated by the motion sensor 148 and the acceleration information generated by the acceleration sensor 146, thereby determining whether the recording system 100 is in an idle state. For example, if the processing unit 110 determines that the recording system 100 is stationary based on the acceleration information and that objects near the recording system 100 have been stationary for a period of time based on the motion sensing signal, the processing unit 110 determines that the recording system 100 is in an idle state. In some embodiments, if the processing unit 110 determines that the video recording system 100 is in an idle state based on the motion sensing signal and the acceleration information, the processing unit 110 controls the camera module 120 to stop capturing images.

[0058] In some embodiments, the gravity sensor 144 is electrically connected to the processing unit 110 via the microcontroller 142 and senses acceleration information of the recording system 100. In some embodiments, the function of the gravity sensor 144 is similar to that of the acceleration sensor 146 and is not further described here.

[0059] In some embodiments, the microphone array 122 is electrically connected to the processing unit 110 , and when the processing unit 110 controls the camera module 120 to start up, the microphone array 122 is controlled to start up to collect the sound of the scene as the sound source of the recorded image.

[0060] In some embodiments, the charging controller 154 is electrically connected to the processing unit 110, the battery 156, and the charging port 152. In some embodiments, if the charging controller 154 detects that an external power source is charging the battery 156 and / or powering the recording system 100 via the charging port 152, the charging controller 154 transmits charging information to the processing unit 110. In some embodiments, when the recording system 100 is powered by an external power source, the processing unit 110 controls the network module 130 to transmit recorded images to the cloud server CS. In some embodiments, when the recording system 100 is not powered by an external power source, the processing unit 110 enters a low-power mode, thereby extending the operating time of the recording system 100.

[0061] In some embodiments, the storage device 114 stores a plurality of neural networks, such as a neural network COM and at least one of the neural networks SPE1 to SPE4. In some embodiments, each of the neural networks COM and SPE1 to SPE4 includes an image segmentation neural network model and an object detection neural network model, thereby performing the task of object detection. In some embodiments, the neural network SPE4 further includes an emotion recognition neural network model, thereby performing the task of emotion recognition (e.g., concentration detection). In some embodiments, the basic structure of the image segmentation neural network model can be implemented by a convolutional neural network architecture (e.g., MobileNetV3), a deep neural network architecture, or other suitable neural network architecture, but the present invention is not limited thereto. In some embodiments, the basic structure of the object detection neural network model can be implemented by a convolutional neural network architecture (e.g., YoloV8), a deep neural network architecture, or other suitable neural network architecture, but the present invention is not limited thereto.

[0062] In some embodiments, neural network COM is a pre-trained model, and the training set used to train neural network COM includes a general object training set, thereby enabling neural network COM to support general object detection tasks. In some embodiments, the general objects include vehicles, traffic lights, people, stores, goods, buildings, facilities, animals, plants, tables, chairs, and other objects. In some embodiments, neural networks SPE1-SPE4 are pre-trained models, and the training set used to train neural networks SPE1-SPE4 includes a specific object training set, thereby enabling neural network COM to support specific object detection tasks.

[0063] In some embodiments, the neural network SPE1 supports detection of specific objects related to the road scene, wherein the specific objects related to the road scene include the status of traffic lights (e.g., red light, yellow light, green light, left and right turn arrow lights), the status of the vehicle's turn signal, double yellow lines, zebra crossings, or other specific objects related to the road scene.

[0064] In some other embodiments, the neural network SPE1 may also determine whether a specific event related to road traffic police has occurred, such as turning left at a red light, failing to yield to pedestrians, and the like.

[0065] In some embodiments, the neural network SPE2 supports detection of specific objects related to the shopping scene, where the specific objects related to the shopping scene include store logos, product categories, or other specific objects related to the shopping scene.

[0066] In some embodiments, the neural network SPE3 supports detection of specific objects related to the travel scene, wherein the specific objects related to the travel scene include famous landmarks (e.g., rocks, landscapes, buildings with specific shapes / contours), amusement facilities, natural environments, special animals, swimming pools, tents, temples, forests, or other specific objects related to the travel scene.

[0067] In some other embodiments, the neural network SPE3 may also determine whether a specific event related to the travel scene occurs, such as fireworks, float parades, and the like.

[0068] In some embodiments, the neural network SPE4 supports detection of specific objects related to the conference scene, where the specific objects related to the conference scene include face frames of participants and information related to the face frames, or other specific objects related to the conference scene.

[0069] The above neural network architecture is only used to illustrate the spirit of the present disclosure and is not intended to limit the present invention.

[0070] In some embodiments, the memory device 112 stores at least one command code. In some embodiments, the processing unit 110 is electrically connected to the memory device 112. In some embodiments, when the processing unit 110 reads at least one command code from the memory device 112, the processing unit 110 executes the image access method 200 of the video recording system 100.

[0071] See also Figure 2 . Figure 2 FIG. 1 is a flow chart of an image access method 200 of the video recording system 100 according to an embodiment of the present disclosure. In some embodiments, the image access method 200 includes steps S210 , S220 , S230 , S240 , and S250 .

[0072] In step S210, a trigger event is detected. In some embodiments, the trigger event is triggered when the processing unit 110 controls the motion sensor 148 via the microcontroller 142 to detect the motion of a person or object in the vicinity of the recording system 100 and generates motion sensing data DATAn. In other embodiments, the trigger event is triggered when an object located within the recording system 100 moves, such as a vehicle or a person. When the trigger event is triggered, the processing unit 110 executes step S220, which activates the camera module 120 to capture image information.

[0073] In capturing image information in step S220 , the processing unit 110 may capture a plurality of images through the camera module 120 to generate a video, wherein the video includes a plurality of frames Fs.

[0074] In step S230, the user mode is analyzed and determined. In some embodiments, the processing unit 110 uses a neural network COM to perform universal object detection on the plurality of frames Fs to identify at least one universal object (e.g., a vehicle, a person, a building, or a storefront) within the plurality of frames Fs. In some embodiments, the processing unit 110 determines an image scene SCE based on the at least one universal object. In some embodiments, the processing unit 110 determines the image scene SCE based on the universal object's category and the area ratio of its bounding box.

[0075] For example, if a plurality of frames Fs contain a large number of vehicle tags, and the number of such vehicle tags exceeds a certain threshold, the processing unit 110 determines that the image scene SCE is a road scene. For another example, if a plurality of frames Fs contain a large number of shopping bag or product tags, and the number of such storefronts, shopping bags, or product tags exceeds a certain threshold, the processing unit 110 determines that the image scene SCE is a shopping scene. For another example, if a plurality of frames Fs contain facility, building, animal, or plant tags, and the number of such facility, building, animal, or plant tags exceeds a certain threshold, the processing unit 110 determines that the image scene SCE is a travel scene. For another example, if a plurality of frames Fs contain facility, building, animal, or plant tags, and the number of such facility, building, animal, or plant tags exceeds a certain threshold, the processing unit 110 determines that the image scene SCE is a travel scene. For another example, if a plurality of frames Fs contain person, table, and chair tags, and the number of such person, table, and chair tags exceeds a certain threshold, the processing unit 110 determines that the image scene SCE is a meeting scene.

[0076] In some embodiments, when the processing unit 110 locks / selects a user mode (i.e., one of the neural networks SPE1-SPE4 to be loaded) based on the image scene SCE, in some embodiments, if the image scene SCE is a road scene, the processing unit 110 locks the user mode to the driving mode M1 and loads the neural network SPE1, and then performs specific object or specific event detection on the plurality of frames Fs based on the neural network SPE1, as described above.

[0077] In some embodiments, if the image scene SCE is a shopping scene, the processing unit 110 locks the user mode in the shopping mode M2 ​​and loads the neural network SPE2, and then performs specific object or specific event detection on multiple frames Fs according to the neural network SPE2, please refer to the above description.

[0078] In some embodiments, if the image scene SCE is a travel scene, the processing unit 110 locks the user mode in the travel mode M3 and loads the neural network SPE3, and then performs specific object or specific event detection on multiple frames Fs according to the neural network SPE3, please refer to the above description.

[0079] In some embodiments, if the image scene SCE is a conference scene, the processing unit 110 locks the user mode in the conference mode M4 and loads the neural network SPE4, and then performs specific object or specific event detection on multiple frames Fs according to the neural network SPE4, please refer to the above description.

[0080] In step S240, the image is analyzed and recorded. In some embodiments, when the processing unit 110 detects the presence of at least one specific object in a plurality of frames Fs, it assigns at least one tag to the frames in which the at least one specific object and / or specific event appears, and stores the frames with the at least one tag. In some embodiments, when the number of frames in which the at least one specific object appears is one, the processing unit 110 stores the frames as image files. In some embodiments, when the number of frames in which the at least one specific object appears is greater than one, the processing unit 110 edits the video into a dynamic image file or a short video file to include the frames containing the at least one specific object.

[0081] In step S250, the image and analysis data are uploaded to the cloud server CS. In some embodiments, the processing unit 110 controls the network module 130 to connect to the cloud server CS to upload the aforementioned image file, dynamic image file, or short video file to the cloud server CS. The frames of the image file, dynamic image file, or short video file include frames with at least one specific object and / or at least one specific event. In some embodiments, the recording system 100 further includes a user interface (not shown), which can be implemented by a display. In some embodiments, the user interface can display all stored tags to the user.

[0082] See also Figure 3A as well as Figure 3B , Figure 3A as well as Figure 3B This is a flow chart illustrating some steps of an image access method 200 according to one embodiment of the present disclosure. In some embodiments, step S230 includes steps S231-S234. In some embodiments, step S230 includes steps S241-S244. In some embodiments, steps S221, S231-S234, and S241-S244 can be executed by the processing unit 110 accessing instructions and data stored in the memory device 112.

[0083] In step S231 , universal object detection is performed on the video.

[0084] In step S232, it is determined whether at least one common object appears in at least one frame of the video. In some embodiments, the processing unit 110 performs common object detection on multiple frames of the video to identify whether at least one common object appears in the multiple frames. If not, step S231 is executed.

[0085] If yes, then step S233 is executed to determine the image scene according to at least one universal object.

[0086] In step S234 , a user mode is determined according to the image scene.

[0087] In step S241, Figure 2 In one of the user modes, perform specific object detection on the video.

[0088] In step S242, it is determined whether at least one specific object or a specific event appears in the plurality of frames. If not, step S241 is executed.

[0089] If yes, then step S243 is executed to attach a label to one of the frames having at least one specific object or at least one specific event.

[0090] In step S244 , the frame with the tag is stored.

[0091] In some embodiments, the image access method 200 further includes the following steps: Storing the frame with the at least one tag on the cloud server CS.

[0092] In some embodiments, the image access method 200 further includes the following steps: determining an image scene based on the category of the at least one universal object and the area ratio of the bounding box, wherein the image scene is one of a road scene, a shopping scene, a travel scene, and a work scene.

[0093] In some embodiments, the image access method 200 further includes the following steps: when the number of frames in which the at least one specific object appears is 1, storing the frame as an image file; and when the number of frames in which the at least one specific object appears is greater than 1, editing the video into a dynamic image file or a short video file to include frames containing the at least one specific object.

[0094] In some embodiments, the image access method 200 further includes the following steps: All stored tags are displayed to a user via a user interface (e.g., a display). When the user selects at least one tag, one of the image file, dynamic image file, and short video file corresponding to the at least one tag is displayed. In some embodiments, the image file, dynamic image file, and short video file include frames containing the at least one specific object and / or the at least one specific event.

[0095] In some embodiments, the at least one specific object is one of a traffic sign, a natural landmark, an artificial landmark, a product trademark, a product item, a store name, and a vehicle.

[0096] In some embodiments, the at least one specific event includes one of the user's own traffic behavior, surrounding traffic behavior, and exercise behavior.

[0097] See also Figures 4A to 4C , Figures 4A to 4C FIG. 1 is a schematic diagram of a road scene according to an embodiment of the present disclosure. Figures 4A to 4C As shown, assuming that the recording system 100 is used as a driving recorder, the processing unit 110 receives the image captured by the camera module 120 and uses the neural network COM to detect that the image contains common objects such as traffic lights, vehicles, and pedestrians. Based on the objects, the mode is locked in the driving mode M1. In the driving mode M1, the processing unit 110 loads the neural network SPE1 to detect the status of the traffic light (for example, Figure 4A as well as Figure 4B The light in the middle is a green light for turning right. Figure 4C In this way, the processing unit 110 adds labels to the plurality of frames according to the red light label and the corresponding Figure 4C The acceleration information of the frame width (for example, the video recording system 100 detects that the vehicle is in an accelerating state) can be used to determine that the driving behavior is suspected of violating the law. The suspected violation includes a situation where there is no violation but it may be misjudged as a violation due to the current situation. In such a situation, the processing unit 110 Figure 4C The video clips are edited based on the frame size and preset time length (for example, 300 seconds before and after) to preserve the image of the vehicle crossing the stop line at the green light for turning right.

[0098] See also Figures 5A to 5C , Figures 5A to 5C FIG. 1 is a schematic diagram of a shopping scene according to an embodiment of the present disclosure. Figures 5A to 5CAs shown, assuming that the recording system 100 is used as a handheld video recorder, the processing unit 110 receives images captured by the camera module 120 and uses the neural network COM to detect that the images contain common objects such as storefronts and merchandise. Based on these objects, the processing unit 110 locks the mode into shopping mode M2. In shopping mode M2, the processing unit 110 loads the neural network SPE2 to detect brands, handbags, luggage, and other merchandise objects in the images and organizes the labels of the brands, handbags, luggage, and other merchandise objects into a product list. In some embodiments, the processing unit 110 may further analyze the brand symbols on the products and mark the brand source on the bounding box of the corresponding product.

[0099] See also Figures 6A to 6B , Figures 6A to 6B FIG. 1 is a schematic diagram of a landscape scene according to an embodiment of the present disclosure. Figures 6A to 6B As shown, assuming that video recording system 100 is used as a handheld video recorder, processing unit 110 receives images captured by camera module 120 and uses neural network COM to detect whether the images contain common objects such as buildings. Based on these objects, the mode is locked in travel mode M3. In travel mode M3, processing unit 110 loads neural network SPE3 to detect whether the buildings in the image are famous landmarks and, based on the detection results, assigns corresponding labels to the video frames.

[0100] In summary, the video recording system 100 disclosed herein is capable of automatically analyzing specific objects in a video and attaching tags to each video frame. This allows users to filter video clips that match specific video scenes (e.g., road scenes, shopping scenes, travel scenes, and meeting scenes) based on the multiple tags. Furthermore, users can filter video clips that match specific objects (e.g., product categories, brand information, natural environment categories, animal categories, facility categories, or famous landmarks) based on the multiple tags. This allows the content of recorded videos to be categorized and more easily searchable, thereby increasing the value of the recorded videos.

[0101] Although the present disclosure has been disclosed above in the form of embodiments, this is not intended to limit the present disclosure. Anyone with ordinary knowledge in the art can make various changes and modifications without departing from the spirit and scope of the present disclosure. Therefore, the scope of protection of the present disclosure should be determined by the appended claims.

Claims

1. A video recording system, characterized in that: Include: a camera module; a memory device storing at least one instruction code; and A processing unit is electrically connected to the camera module and the memory device, and when the processing unit reads the at least one instruction code, at least the following steps are performed: Capturing multiple images through the camera module to generate a video; identifying at least one common object in a plurality of frames in the video; determining an image scene according to the at least one universal object; According to the image scene, performing specific object detection on the plurality of frames to identify whether at least one specific object or at least one specific event appears in the plurality of frames; as well as A label is attached to at least one of the frames having the at least one specific object or the at least one specific event.

2. The video recording system according to claim 1, wherein: Also includes: A storage device is electrically connected to the processing unit and stores a first neural network and a plurality of second neural networks, wherein the processing unit further performs the following steps: Performing common object detection on the plurality of frames in the video using the first neural network to identify the at least one common object in the plurality of frames; determining the image scene according to the at least one universal object; selecting one of the plurality of second neural networks according to the image scene; as well as The one of the plurality of second neural networks is utilized to perform specific object detection on the plurality of frames to identify the at least one specific object or the at least one specific event in the plurality of frames.

3. The video recording system according to claim 1, wherein: Also includes: a motion sensor electrically connected to the processing unit and sensing an ambient environment to generate a motion sensing signal; When the processing unit receives the motion sensing signal, it controls the camera module to start up to generate the video.

4. The video recording system according to claim 3, wherein: Also includes: an acceleration sensor electrically connected to the processing unit and sensing acceleration information of the camera module; The processing unit controls the camera module to stop shooting according to the acceleration information and the motion sensing signal.

5. The video recording system according to claim 4, wherein: Also includes: A network module is electrically connected to the processing unit and transmits data to a cloud server. The processing unit transmits the at least one frame with the tag to the cloud server according to the acceleration information and the motion sensing signal.

6. An image access method, characterized in that: Include: Performing common object detection on a plurality of frames in a video to identify at least one common object in the plurality of frames; determining an image scene according to the at least one universal object; According to the image scene, performing specific object detection on the plurality of frames to identify whether there is at least one specific object or at least one specific event in the plurality of frames; When it is detected that the at least one specific object appears in the plurality of frames, at least one label is attached to the plurality of frames in which the at least one specific object or the at least one specific event appears; as well as The frames having the at least one tag are stored.

7. The image access method according to claim 6, wherein: Also includes: The plurality of frames having the at least one tag are stored on a cloud server.

8. The image access method according to claim 6, wherein: Also includes: Determine the image scene based on the category of the at least one common object and the area ratio of the bounding box, The image scene is one of a road scene, a shopping scene, a travel scene and a work scene.

9. The image access method according to claim 6, wherein: Also includes: When the number of the frames in which the at least one specific object appears is 1, storing the frame as an image file; as well as When the number of the frames in which the at least one specific object appears is greater than 1, the video is edited into a dynamic image file or a short video file to include the frames containing the at least one specific object.

10. The image access method according to claim 6, wherein: Also includes: displaying all stored at least one tag to a user through a user interface; When the user clicks on the at least one tag, one of an image file, a dynamic image file, and a short video file corresponding to the at least one tag is displayed. The image file, the dynamic image file and the short video file include the plurality of frames having the at least one specific object or the at least one specific event.

11. The image access method according to claim 6, wherein: The at least one specific object is one of a traffic sign, a natural landmark, an artificial landmark, a product trademark, a product item, a store name, and a vehicle.

12. The image access method according to claim 6, wherein: The at least one specific event includes one of a own traffic behavior, a surrounding traffic behavior, and a sports behavior.

13. A computer-readable medium, characterized in that The computer-readable medium stores at least one instruction code. When the at least one instruction code is read by a processing unit, the processing unit is caused to execute the following steps: Performing common object detection on a plurality of frames in a video to identify at least one common object in the plurality of frames; determining an image scene according to the at least one universal object; According to the image scene, performing specific object detection on the plurality of frames to identify whether there is at least one specific object or at least one specific event in the plurality of frames; When it is detected that the at least one specific object appears in the plurality of frames, at least one label is attached to the plurality of frames in which the at least one specific object or the at least one specific event appears; as well as The frames having the at least one tag are stored.

14. The computer-readable medium of claim 13, wherein: When the at least one instruction code is read by the processing unit, the processing unit is further caused to execute the following steps: Connect to a cloud server; and The plurality of frames having the at least one tag are stored in the cloud server.

Images