专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明公开

EP4390858A1 MODEL TRAINING METHOD AND DEVICE, AND TARGET DETECTION METHOD AND DEVICE 审中-实审
公开(公告)号：EP4390858A1
公开(公告)日：2024-06-26
申请号：EP23838710.4
申请日：2023-06-29
申请人： Mashang Consumer Finance Co., Ltd.
发明人： LV, Yongchun , ZHU, Hui , WANG, Yu , ZHOU, Xunyi , ZENG, Dingheng , JIANG, Ning
IPC分类号： G06V10/25
CPC分类号： G06V10/44 , G06V10/766 , G06V10/75 , G06V10/25
摘要： The present application provides a model training method, a target detection method and a device. In a model training phase, based on a first reference bounding box and its corresponding real bounding box and real category, a model to be trained is prompted to continuously learn a bounding box distribution and an identifying of a category of a target object; and determine a loss value based on a determination result set output by a determination sub-model, and then continuously update model parameters for multiple rounds of iterations based on the loss value, where the determination result set includes a first determination result that represents a bounding box distribution similarity, and a second determination result that represents a category similarity between the first prediction category and the real category.

2. 发明公开

EP4345645A1 USER QUESTION LABELING METHOD AND DEVICE 审中-实审
公开(公告)号：EP4345645A1
公开(公告)日：2024-04-03
申请号：EP22913488.7
申请日：2022-09-07
申请人： Mashang Consumer Finance Co., Ltd.
发明人： GENG, Fuming , WU, Haiying , JIANG, Ning
IPC分类号： G06F16/332
摘要： Embodiments of the present application provide a user question labeling method and apparatus. By obtaining dialogue data generated by a manual customer service system, determining a target quick response statement corresponding to a target raised question, and then determining, in combination with a correspondence between a preset quick response statement and a user intention label, a target intention label corresponding to the target quick response statement, so as to obtain a correspondence between the target raised question and the target intention label, labeling user intention for a user question is thus achieved and then a generated user intention recognition sample set is applied to a user intention recognition model in an intelligent customer service system. Not only can the labeling efficiency of the user question be improved, so that a large number of user intention recognition samples are provided for the intelligent customer service system, but also the coverage rate of the user questions that cannot be recognized by the intelligent customer service system included in the user intention recognition sample set can be improved. Therefore, recognition accuracy of the user intention recognition model is improved, and service quality of the intelligent customer service system is thus ensured.

3. 发明公开

EP4344199A1 SPEECH AND IMAGE SYNCHRONIZATION MEASUREMENT METHOD AND APPARATUS, AND MODEL TRAINING METHOD AND APPARATUS 审中-实审
公开(公告)号：EP4344199A1
公开(公告)日：2024-03-27
申请号：EP22866437.1
申请日：2022-08-25
申请人： Mashang Consumer Finance Co., Ltd.
发明人： WANG, Chun , ZENG, Dingheng , WU, Haiying , ZHOU, Xunyi , JIANG, Ning
IPC分类号： H04N17/00 , H04N21/439 , H04N21/44 , H04N21/43 , G06N3/04 , G06N3/08 , G06V20/40
摘要： The application provides a method and an apparatus for measuring speech-image synchronicity, and a method and an apparatus for training a model, where the method for measuring speech-image synchronicity includes: acquiring a speech segment and an image segment of a video, where there is a correspondence between the speech segment and the image segment in the video; performing any of following operations: converting the speech segment into a specific signal and acquiring, through a pre-trained model for measuring speech-image synchronicity, a speech feature of the specific signal and a visual feature of the image segment, where the specific signal is independent of a personal feature of a speaker in the speech segment; or, generating a contour map of a target character according to the image segment and acquiring, through a pre-trained model for measuring speech-image synchronicity, a visual feature of the contour map and a speech feature of the speech segment, where the contour map is independent of a personal feature of the target character; or, converting the speech segment into a specific signal, generating a contour map of a target character according to the image segment, and acquiring, through a pre-trained model for measuring speech-image synchronicity, a speech feature of the specific signal and a visual feature of the contour map; and determining, according to the speech feature and the visual feature, whether there is synchronicity between the speech segment and the image segment, where the synchronicity is used for characterizing matching between a sound in the speech segment and a movement of a target character in the image segment.

4. 发明公开

EP4386692A1 TRAINING METHOD FOR IMAGE CLASSIFICATION MODEL, IMAGE CLASSIFICATION METHOD, AND RELATED DEVICE 审中-实审
公开(公告)号：EP4386692A1
公开(公告)日：2024-06-19
申请号：EP23842011.1
申请日：2023-06-26
申请人： Mashang Consumer Finance Co., Ltd.
发明人： LV, Yongchun , ZHU, Hui , ZHOU, Xunyi , JIANG, Ning , WU, Haiying
IPC分类号： G06V10/764
CPC分类号： G06V10/764 , G06V10/774
摘要： The present application discloses a method for training an image classification model, an image classification method and related device. The method includes: obtaining an image set; discriminating a labeled image and an unlabeled image by a target image classification sub-model of the image classification model, and determining first classification reference information of the labeled image and first classification reference information of the unlabeled image; discriminating the unlabeled image by a non-target image classification sub-module of the image classification model, and determining second classification reference information of the unlabeled image; determining a classification loss of the target image classification sub-model according to the first classification reference information of the labeled image, the category label of the labeled image, and the second classification reference information of the unlabeled image; tuning a model parameter of the image classification model according to the classification loss of the first image classification sub-model and the classification loss of the second image classification sub-model.

5. 发明公开

EP4379669A1 PROCESSING METHOD FOR IMAGE RECOGNITION MODEL, AND RELATED PRODUCT 审中-实审
公开(公告)号：EP4379669A1
公开(公告)日：2024-06-05
申请号：EP23849243.3
申请日：2023-07-26
申请人： Mashang Consumer Finance Co., Ltd.
发明人： LV, Yongchun , ZHU, Hui , ZHOU, Xunyi , JIANG, Ning , WU, Haiying
IPC分类号： G06V10/25
CPC分类号： G06V10/82 , G06V10/25 , G06V10/764 , G06V10/774
摘要： Embodiments of the present application provide a processing method for an image recognition model and a related product. The processing method for the image recognition model includes: obtaining an image sample; determining a target object positioning box of the image sample, where the target object positioning box covers an area where an effective feature of the image sample is located; adjusting a pixel value of at least one pixel within the target object positioning box to obtain a preprocessed image sample; and training a network model based on the preprocessed image sample to obtain the image recognition model. The present application improves the concentration of the effective feature in the sample image and thus improves the regularization effect and reduces overfitting of the network model.

6. 发明公开

EP4394723A1 LABEL GENERATION METHOD AND APPARATUS, IMAGE CLASSIFICATION MODEL METHOD AND APPARATUS, AND IMAGE CLASSIFICATION METHOD AND APPARATUS 审中-实审
公开(公告)号：EP4394723A1
公开(公告)日：2024-07-03
申请号：EP23842015.2
申请日：2023-06-26
申请人： Mashang Consumer Finance Co., Ltd.
发明人： LV, Yongchun , ZHU, Hui , ZHOU, Xunyi , JIANG, Ning , WU, Haiying
IPC分类号： G06V10/44 , G06V10/764 , G06V10/772 , G06V10/82 , G06N3/04 , G06N3/08
CPC分类号： G06V10/82 , G06N3/04 , G06V10/44 , G06N3/096 , G06N3/08 , G06V10/764 , G06N3/0895 , G06V10/772
摘要： The present application discloses a label generation method, an image classification model, an image classification method and a device. The label generation method includes: classifying and identifying an unlabeled image by a teacher model and a student model respectively to obtain classification reference information of the teacher model and classification reference information of the student model; the classification reference information of each model includes N detection objects included in the unlabeled image and a probability of each detection object in each category of a plurality of categories; determining a number of detection objects of each category in the teacher model according to the classification reference information of the teacher model, and determining a number of detection objects of each category in student model according to the classification reference information of the student model; determining a probability threshold of each category according to the number of detection objects of each category in the teacher model and the number of detection objects in the student model; generating a pseudo label of the unlabeled image for the student model according to the second classification reference information and the probability threshold of each predicted category.

7. 发明公开

EP4345771A1 INFORMATION PROCESSING METHOD AND APPARATUS, AND COMPUTER DEVICE AND STORAGE MEDIUM 审中-实审
公开(公告)号：EP4345771A1
公开(公告)日：2024-04-03
申请号：EP22936416.1
申请日：2022-12-30
申请人： Mashang Consumer Finance Co., Ltd.
发明人： WANG, Chun , ZENG, Dingheng , ZHOU, Xunyi , JIANG, Ning
IPC分类号： G06V20/40
摘要： The present application discloses an information processing method, a device, a computer device and a storage medium. The embodiment of the present application obtains a target video; inputs a face image of each frame into a first target preset backbone model, and outputs a first target image feature; obtains a first target identity coefficient and a first target texture coefficient of the face image corresponding to the first target image feature in a different frame of the target video; inputs the first target identity coefficient into a second target preset backbone model and outputs a first target identity feature; inputs the first target texture coefficient into a third target preset backbone model and outputs a first target texture feature; splices the first target image feature, the first target identity feature and the first target texture feature to obtain a first target feature; inputs the first target feature into a target preset head network model, and outputs a first target expression coefficient. Combining prior knowledge in a video to predict the expression coefficient greatly improves an accuracy of the expression coefficient that is extracted.

8. 发明公开

EP4345770A1 INFORMATION PROCESSING METHOD AND APPARATUS, COMPUTER DEVICE, AND STORAGE MEDIUM 审中-实审
公开(公告)号：EP4345770A1
公开(公告)日：2024-04-03
申请号：EP22936399.9
申请日：2022-12-19
申请人： Mashang Consumer Finance Co., Ltd.
发明人： WANG, Chun , ZENG, Dingheng , ZHOU, Xunyi , JIANG, Ning
IPC分类号： G06V20/40 , G06N3/04
摘要： The embodiment of the present application provides an information processing method, a device, a computer device, and a storage medium. The embodiment of the present application obtains a target video, each frame of the target video including a face image corresponding to a same object; inputting the face image of each frame into a first target preset backbone model, and outputting a first target image feature; inputting the first target image feature into a first target preset network model, and correspondingly outputting a first target expression coefficients; among them, the first target preset network model is obtained by obtaining a shared identity coefficient and a shared texture coefficient, and performing a corresponding prediction optimization training on an expression coefficient. By decoupling the first target expression coefficient from non-expression coefficient, an accuracy of the expression coefficient that is extracted is greatly improved.

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式