40results about How to "Annotation implementation" patented technology

The invention discloses a comment display method, device and equipment and a storage medium and relates to the field of interactive smart laptops. The comment display method specifically includes: receiving a screen protection instruction; displaying first screen projection content; receiving a first comment signal which is used for commenting the first screen projection content; displaying a first comment trail which is generated according to the first comment signal; receiving a forward page turning instruction which is used for controlling forward page turning of the first screen projectioncontent; displaying second screen projection content after page turning; receiving a backward page turning instruction which is used for controlling backward page turning of the second screen projection content; displaying the first screen projection content and the first comment trail. By adopting the technical scheme, the technical problem that equipment for displaying screen projection data cannot realize commenting on screen projection content in the prior art can be solved, commenting on the screen projection content is realized, and synchronous appearing or disappearing of comment and the screen protection content is ensured.

The invention discloses a live broadcast annotation method and device, a terminal and a live broadcast system. The live broadcast annotation method comprises the steps of acquiring first touch path information corresponding to a touch operation performed by a local user on a local touch screen; drawing a first annotation track corresponding to the touch operation according to the first touch path information, sending the first touch path information to a display terminal which establishes connection with the local terminal so as to enable the display terminal to draw a second annotation track displayed on the display terminal according to the first touch path information; and displaying the first annotation track to the local user through a screen. Through adopting the above technical scheme, the embodiment of the invention can perform annotation on the live broadcast content in the live broadcast process, so that users of the display terminal are enabled to be more targeted when watching the live broadcast, and the display effect of the live broadcast is improved.

The embodiment of the invention provides a video annotation method and device, a tracking method and device, computer equipment and a storage medium. The video annotation method comprises the steps ofdisplaying an inter-frame association relationship between a to-be-annotated video frame and a to-be-annotated video to which the to-be-annotated video frame belongs; receiving first input information used for indicating a labeling result of a to-be-labeled video frame and second input information used for indicating a video frame range using the labeling result, wherein the video frame range isdetermined according to an inter-frame association relationship of the to-be-labeled video; obtaining a labeling result of a to-be-labeled video frame according to the first input information, and obtaining a labeling result of a video frame in a video frame range according to the second input information; respectively calculating mapping matrixes of a previous video frame and a next video frame in a video; multiplying the two mapping matrixes to obtain an inter-frame mapping matrix, wherein the inter-frame mapping matrix reflects the inter-frame offset of the object image between the previousvideo frame and the next video frame.

Disclosed are a spliced display system whole-wall labeling method and system, and a spliced display system. The method comprises: respectively acquiring positions of a signal window and desktop labeling information on a spliced wall; according to the position of the signal window and the position of the desktop labeling information, computing the position of an overlapping portion of the signal window and the desktop labeling information, and computing labeling information corresponding to the overlapping portion; and according to the position of the overlapping portion, displaying the labeling information corresponding to the overlapping portion on the signal window accordingly, and displaying the signal window and the desktop labeling information on the spliced wall. The method and system according the present invention implement whole-wall labeling, and avoid the circumstance that labeling content is shaded when whole-wall labeling is conducted in the spliced display system, thereby improving user experience.

The invention provides a sample labeling method and a computer storage medium. The sample labeling method comprises the following steps: carrying out character detection on a to-be-labeled sample image through a character detection model, and obtaining a character box; determining the average height of the character boxes according to the number of the character boxes and the height of each character box; screening out at least one first candidate character box from all character boxes according to the average height of the character boxes; for each first candidate character box, determining the character box with the minimum distance from the current first candidate character box as a current second candidate character box in the vertical direction, and determining a reference character box in the horizontal direction; if the horizontal center line of the current second candidate character box passes through the reference character box in the horizontal direction, determining the current first candidate character box and the current second candidate character box as candidate character boxes to be merged; and generating labeling information of the sample image according to the processing of the candidate character boxes to be merged. For the sample labeling method, the marking accuracy of the sample marking method is better.

The invention discloses a simultaneous multi-screen marking method and device. The method includes: registering multiple display screens to be marked; acquiring marking strokes and distributing the marking strokes to the display screens to be marked; marking displayed images of the display screens according to the marking strokes on the display screens to be marked. The method allows the displayed images of the registered multiple display screens to be marked at the same time. The device comprises a marking server, a marking stroke acquisition module, a marking stroke distribution module and a stroke displaying module. The device is a hardware base for the implementation of the method. The use of the method with the device allows the displayed images on the registered multiple display screens to be marked at the same time.

The invention discloses an image management display method and device, computer equipment and a storage medium, and the method is based on the field of artificial intelligence, and comprises the following steps: creating a database which is used for storing an X-ray image and is provided with a query tree, wherein the query tree is correlated with the file information of the X-ray image; calling the target X-ray image from the database through the query tree, obtaining file information of the target X-ray image, and displaying the target X-ray image and the file information thereof; receivingand storing line information generated by drawing lines on the target X-ray image by a client, and closing the line information to generate a closed curve; receiving a predefined label input by the client according to the closed curve, and associating the predefined label with the closed curve; and associating the predefined label with the query tree to obtain a labeled X-ray image, and storing the labeled X-ray image in the database. According to the invention, other doctors can visually and accurately obtain the lesion area and perform pathological judgment on the lesion area, and illness state communication obstacles generated between the mammary gland doctor and other doctors are eliminated.

The invention discloses an image labeling device, which comprises an image block determining unit, a feature information determining unit and a labeling unit, wherein the image block determining unitis used for determining at least one image block in an image to be labeled; the feature information determining unit is used for independently determining the feature information of each image block,wherein the feature information is used for uniquely expressing the corresponding image block; and the labeling unit is used for labeling the image block corresponding to target feature information soas to label a labeling object in the image to be labeled when the target feature information matched with datum feature information is in the presence in each piece of feature information, wherein the datum feature information corresponds to the labeling object. By use of the image labeling way, the image block is automatically labeled so as to overcome the problem of low manual labeling efficiency in the prior art, and a foundation is laid for effectively obtaining a training sample.

The invention provides a data labeling method and device and data processing equipment, and the method comprises the steps: carrying out the at least one iteration processing of a classification model, so as to enable the accuracy of the classification model to meet a preset condition; and processing at least one part of the plurality of pieces of to-be-labeled data by utilizing the obtained classification model to obtain an automatic labeling result, wherein each iterative processing comprises the following steps of: respectively inputting other to-be-labeled data, except a target data set, in a plurality of pieces of to-be-labeled data into the classification model to obtain a classification result; selecting at least part of the to-be-labeled data of which the confidence coefficient ofthe classification result is within a preset range from the other to-be-labeled data, and adding the selected to-be-labeled data into a target data set; and training a classification module accordingto the manual labeling result of the to-be-annotated data in the target data set. Therefore, automatic labeling of batch data can be realized under the condition of improving the data labeling quality.

The invention provides a method and device for time alignment calibration, event annotation and database generation. The time alignment calibration method comprises the following steps: (A) acquiringan event stream of a target object shot by a dynamic vision sensor and a video image of the target object shot by an auxiliary vision sensor; (B) determining a key frame for representing obvious movement of the target object from a video image; (C) respectively mapping an effective pixel location of the target object in the key frame and an effective pixel location of the target object in the adjacent frame of the key frame to an imaging plane of the dynamic vision sensor, thereby forming multiple target object templates; (D) determining a first target object template covering the most eventsin the first event stream section from multiple target object templates; and (E) taking the time alignment relation of the middle moment of the first event stream section and the timestamp of the frame corresponding to the first target object template as the time alignment relation between the dynamic vision sensor and the auxiliary vision sensor.

The invention provides an image labeling method, device and system and a cell labeling method, and relates to the field of artificial intelligence. The method comprises the steps of obtaining a to-be-labeled image, wherein the to-be-labeled image comprises an identifier for pre-labeling a target object; performing image segmentation on the to-be-labeled image to obtain superpixels corresponding tothe to-be-labeled image; and performing super-pixel clustering according to the pixel information corresponding to the identifier and the pixel information in the super-pixels so as to mark the target object. According to the method and the apparatus, the target object is prevented from being completely labeled in a manual labeling manner, so that manpower and material resources are saved and thelabeling efficiency is improved; in addition, image segmentation is carried out on the to-be-labeled image to form superpixels, the calculation amount during pixel clustering can be reduced, and thelabeling efficiency is further improved.

The invention provides a chat corpus labeling method, which comprises the steps of obtaining a question text set matched with chat corpus, the question text set comprising at least one question text which does not obtain a corresponding reply statement; expanding question texts in the question text set through a question text expansion model network in a chat corpus annotation model to obtain corresponding question text pairs; in response to the obtained question text pair, determining a reply statement corresponding to the question text in the question text set through a question-answer modelnetwork in the chat corpus annotation model; and correcting the question text pair and the reply statement, and establishing association between the question text pair and the reply statement. The invention further provides a chat corpus labeling device, electronic equipment and a storage medium. The chatting corpus annotation method and the device can achieve annotation of the chatting corpus.

The invention discloses a method and a device for detecting a text in an image, and relates to the technical field of image processing. A specific embodiment of the method comprises the steps of determining a labeling template in a labeling image and a text line in the labeling template; based on the shape of the labeling template contour, labeling the characters in all the text lines by adoptingat least one polygonal labeling box to obtain a character labeling sample; training the text detection model based on the text annotation sample to obtain a trained text detection model, the input ofthe trained text detection model being an image, and the output being the position of the Chinese row of the image; and predicting the text lines in the image to be detected by adopting the trained text detection model to obtain the positions of the text lines output by the trained text detection model. According to the embodiment, the text detection efficiency is improved.

Embodiments of the invention provide a method and device for WeChat annotation. The method comprises: monitoring a started WeChat service program, determining contents shown by the WeChat service program in an effectively display region; obtaining an annotation position on the shown content, and obtaining and loading input annotation information on the content corresponding to the annotation position. In the method, firstly, the started WeChat service program is monitored, the contents shown by the WeChat service program in the effectively display region is determined, and then, the annotation position on the shown content is obtained, and the input annotation information is loaded on the content corresponding to the annotation position, so that annotation on the contents shown in the WeChat is realized.

The embodiment of the invention discloses a labeling method and device, a server and a storage medium. The method comprises the following steps of acquiring a to-be-labeled object, a labeled object atthe previous moment and a prediction model at the previous moment; updating the prediction model at the previous moment based on the labeled object at the previous moment to obtain a prediction modelat the current moment; performing association type prediction processing on the to-be-labeled object by adopting the prediction model at the current moment to obtain a relationship among a pluralityof types associated with the to-be-labeled object; and labeling the to-be-labeled object according to the relationship among the multiple types associated with the to-be-labeled object to obtain the labeled object at the current moment, thereby realizing labeling of the to-be-labeled object. Before each classification, the prediction model can be updated according to the last classification result, so that the prediction model can be self-iterated and optimized, the prediction effect of the prediction model on the current to-be-labeled object is accurate, and the labeling accuracy can be improved.

The embodiment of the invention relates to the technical field of machine vision, and discloses an article labeling method and device, an electronic device and a storage medium. The method comprises the following steps: generating a binary article labeling model for labeling a target object and a non-target object; inputting a plurality of to-be-labeled pictures; wherein the target objects in theto-be-labeled picture are labeled with categories, and each target object in the to-be-labeled picture containing the target object is one type; and automatically labeling the to-be-labeled picture through the binary article labeling model to obtain position information of the target object, and taking the labeled category as the category of the target object output by the binary article labelingmodel. According to the embodiment of the invention, article labeling can be realized more quickly with low cost.

The invention relates to a medical image labeling system. The medical image annotation system comprises an image acquisition module, a display module and an annotation module. A three-dimensional medical image is obtained through the image acquisition module, and then two-dimensional images of a coronal plane, a sagittal plane and a transverse plane of a three-dimensional medical image can be respectively displayed in different areas of a display interface by utilizing the display module; the annotation module can be used for obtaining the position surface annotation position of a user on thetwo-dimensional image of any position surface in the display interface, and the position surface annotation position corresponds to the three-dimensional medical image to obtain the three-dimensionalannotation position; the display module can be used for displaying the position information of the three-dimensional annotation position corresponding to each position surface on each position surfacein the display interface. By means of the system, the annotation of the specific position in the three-dimensional medical image is achieved, so convenience is achieved and high annotation efficiencyis achieved; in addition, corresponding voxels can be obtained based on the image data of the bit plane annotation position, and detailed image information of the three-dimensional annotation position can be obtained conveniently.

The invention discloses a human body key point detection method and system based on a Wi-Fi signal. The method comprises the following steps: recording CSI data and CSI timestamps in signal coverage ranges of a Wi-Fi signal transmitting end and a Wi-Fi signal receiving end; and recording video data and UNIX timestamps; performing cross-modal indirect labeling on the CSI data by using the collected video data; taking the marked CSI data as training data, inputting the training data into a Wi-Fi Transformer model for training, and storing the training data after the Wi-Fi Transformer model converges; and inputting the Wi-Fi data to be detected into the trained Wi-Fi Transformer model, and outputting the key point data of the human body. According to the method and system, the Wi-Fi Transformer model is adopted to detect the key points of the human body, and the method is simple, rapid and accurate; sensitive privacy such as human faces cannot be exposed; therefore, the method and system can be used in the fields of daily life and health medical treatment.

The invention provides a voice data labeling method and device, electronic equipment and a computer readable medium. The method comprises the following steps: receiving voice data to be labeled, and performing voice recognition on the voice data to obtain a recognition text; obtaining a user confirmation text after the user confirms the recognition text; extracting automatic labeling features fromthe recognition text and the user confirmation text; and labeling the voice data according to the automatic labeling features and a pre-constructed automatic labeling model. Due to the fact that theautomatic annotation features are extracted, the voice data are annotated according to the automatic annotation features and the automatic annotation model, automatic annotation of the voice data canbe achieved, manual annotation is not needed, and therefore the problems existing in manual annotation can be solved, the voice data annotation efficiency is improved, and the cost is reduced.

The invention discloses a method for labeling an audio by using a deep learning model. The method comprises the following steps: A, acquiring the audio and performing voice preprocessing on the acquired audio; b, inputting the audio data subjected to voice preprocessing into a deep learning model for voice recognition and voice annotation, and labeling the audio according to the voice annotation,wherein the deep learning model comprises a deep neural network and a long-short-term memory unit; and C, performing manual proofreading on the label output by the deep learning model. According to the method disclosed by the invention, the tedious work of manual listening, manual labeling and manual proofreading is converted into the work of only needing manual proofreading, and other operationsare automatically carried out by the system model, so that the manpower and time cost can be greatly saved, and the effectiveness is guaranteed.

The invention provides a sample labeling method and a computer storage medium. The sample labeling method comprises the steps: obtaining a sample image; performing character detection on the sample image through a character detection model to obtain a character frame of the character position of each character in the sample image; determining the average width of the character frames according tothe number of the character frames and the width of each character frame; screening out at least one first candidate character box according to the average width of the character boxes; determining acharacter box with the minimum distance from the current first candidate character box as a current second candidate character box in the horizontal direction, and determining a character box with theminimum distance from the current first candidate character box and the width greater than a first reference value as a reference character box in the vertical direction; if the vertical center lineof the current second candidate character box in the vertical direction passes through the reference character box, determining the current first candidate character box and the current second candidate character box as candidate character boxes to be combined; and generating labeling information of the sample image according to the processing of the candidate character boxes to be merged.

The embodiment of the invention provides a video annotation method and device, and belongs to the technical field of computers, and the method comprises the steps: obtaining the annotation informationof a target object before the annotation of the target object in a to-be-annotated video, and enabling the annotation information to comprise a plurality of images and the position information of thetarget object in each image, wherein the plurality of images include the target object shot at different shooting angles; based on the annotation information, training a preset annotation model to obtain a first annotation model capable of annotating the target object; and labeling the target object in the to-be-labeled video through the first labeling model to obtain a labeled video. By adoptingthe technical scheme provided by the embodiment of the invention, the problem of poor generalization of video annotation can be solved.

The invention discloses a method for enabling the function of a magnifying glass, which is applied to an electronic device. The electronic device has an image acquisition unit. The method includes: when the electronic device is in a lock screen state or a use state, acquiring A first distance value of the distance between an object in front of the image acquisition unit and the image acquisition unit; judging whether the first distance value satisfies a predetermined condition, and generating a judgment result; when the judgment result indicates that the first When a distance satisfies the predetermined condition, a first opening instruction is generated; and the first opening instruction is executed to control the electronic device to open the function of the magnifying glass.

The invention provides an attack data set malicious fragment labeling method and system based on LSTM. The method comprises the following steps: extracting key values of all groups of parameters in amalicious URL; converting the key values into feature representations; inputting the feature representations into a trained LSTM model, and predicting to obtain a prediction result value with the maximum numerical value; and obtaining a corresponding malicious fragment according to the prediction result value with the maximum numerical value, obtaining the position of the malicious fragment in themalicious URL, and labeling the malicious fragment. The system comprises: a data processing unit used for extracting key values of all groups of parameters in the malicious URL and converting the keyvalues into the feature representations; a data prediction unit used for inputting the feature representations into the trained LSTM model for prediction to obtain the prediction result value with the maximum numerical value; and a data labeling unit used for acquiring the corresponding malicious fragment according to the prediction result value with the maximum numerical value, thereby acquiringthe position of the malicious fragment in the malicious URL and labeling the malicious fragment.

The embodiment of the present invention relates to the technical field of machine vision, and discloses an item labeling method and device, electronic equipment, and a storage medium. The method includes: generating a two-category item labeling model for labeling targets and non-targets; inputting a number of pictures to be marked; wherein, the targets in the pictures to be marked are marked with categories, and each target containing the target The target object in the marked image is one type; the image to be marked is automatically marked through the binary classification item labeling model to obtain the position information of the target object, and the marked category is used as the target object category output by the binary classification item labeling model. The embodiment of the present invention can realize item labeling more quickly and at low cost.

The invention discloses a comment display method, device, equipment and storage medium, and relates to the field of interactive intelligent flat panel, which specifically includes: receiving a projection screen instruction; displaying the content of the first projection screen; receiving the first comment signal, and using the first comment signal For annotating the content of the first screen projection; displaying the first annotation track, the first annotation track is generated according to the first annotation signal; receiving a forward page turning instruction, which is used to control the direction of the first projection screen content to Turn the page forward; display the content of the second projected screen after the page is turned; receive the command to turn the page backward, which is used to control the content of the second projected screen to turn backward; display the content of the first projected screen and the first comment track. Adopting the above technical solution can solve the technical problem that the equipment displaying projection data in the prior art cannot annotate the projection content, realize the annotation for the projection content, and at the same time ensure that the annotation and the projection content appear or disappear synchronously.

The invention provides a medical image labeling method, a medium and electronic equipment. The medical image labeling method comprises the steps of obtaining seed pixel points of a to-be-labeled object; determining representative pixel points according to the CT values of the pixel points in the three-dimensional neighborhood of the seed pixel points; and labeling the object to be labeled according to the representative pixel points. The medical image labeling method can realize three-dimensional labeling of the to-be-labeled object.

The embodiment of the invention provides a picture labeling method, device and equipment. The method comprises the steps of obtaining a label relationship of a preset number of picture pairs in a picture data set, wherein the label relationship is used for representing whether the picture pairs have the same label; determining a label relationship set of the picture data set according to the labelrelationship of the preset number of picture pairs and the transmissibility of the label relationship; and determining the label of each picture in the picture data set according to the label relationship set. According to the method, the picture labeling speed and accuracy are improved, and the problems that an existing picture labeling method consumes time and labor and is low in accuracy are solved.