26 September 2023 in IEEE Transactions on Circuits and Systems for Video Technology
Blur artifacts can seriously degrade the visual quality of images, and numerous deblurring methods have been proposed for specific scenarios. However, in most real-world images, blur is caused by different factors, e.g ., motion, and defocus. In this paper, we address how other deblurring methods perform in the case of multiple types of blur. For in-depth performance evaluation, we construct a new large-scale multi-cause image deblurring dataset (MC-Blur), including real-world and synthesized blurry images with different blur factors. The images in the proposed MC-Blur dataset are collected using other techniques: averaging sharp images captured by a 1000-fps high-speed camera, convolving Ultra-High-Definition (UHD) sharp images with large-size kernels, adding defocus to images, and real-world blurry images captured by various camera models. Based on the MC-Blur dataset, we conduct extensive bench-marking studies to compare SOTA methods in different scenarios, analyze their efficiency, and investigate the buildataset’s capacity. These benchmarking results provide a comprehensive overview of the advantages and limitations of current deblurring methods, revealing our dataset’s advances. The dataset is available to the public at https://github.com/HDCVLab/MC-Blur-Dataset.
11 April 2025 in Pattern Recognition
2 May 2024 in IEEE Access
Most video anomaly detection approaches are based on non-semantic features, which are not interpretable, and prevent the identification of anomaly causes. Therefore, we propose a caption-guided interpretable video anomaly detection framework that explains the prediction results based on video captions (semantic). It utilizes non-semantic features to fit the dataset and semantic features to provide common sense and interpretability to the model. It automatically stores representative anomaly prototypes and uses them to guide the model based on similarity with these prototypes. Specifically, we use video memory to represent the content of videos, which includes video features (non-semantic) and caption information (semantic). The proposed method generates and updates a memory space during training, and predicts anomaly scores based on the memory similarities between the input video and the stored memories. The stored captions can be used as descriptions of representative anomaly actions. The proposed module can be easily integrated with existing methods. The interpretability and reliable detection performance of the proposed method are evaluated through extensive experiments on public benchmark datasets.
1 January 2009 in Electronic Workshops in Computing
This paper presents a system for interaction with a display via hand pointing, where a single CCD camera on top of the screen is directed towards the viewers. An attention mechanism based on face and hand detection allows one user to take control of the interface. Face recognition is used for identification and customisation. The system allows the user to control the screen pointer by tracking their fist. Onscreen items can be selected using one of four activation mechanisms. Current sample applications include browsing image and video collections as well as viewing a gallery of 3D objects. In experiments we demonstrate the performance of the vision components in challenging conditions.
14 February 2022 in arXiv
E-commerce platforms generate vast amounts of customer behavior data, such as clicks and purchases, from millions of unique users every day. However, effectively using this data for behavior understanding tasks is challenging because there are usually not enough labels to learn from all users in a supervised manner. This paper extends the BERT model to e-commerce user data for pre-training representations in a self-supervised manner. By viewing user actions in sequences as analogous to words in sentences, we extend the existing BERT model to user behavior data. Further, our model adopts a unified structure to simultaneously learn from long-term and short-term user behavior, as well as user attributes. We propose methods for the tokenization of different types of user behavior sequences, the generation of input representation vectors, and a novel pretext task to enable the pre-trained model to learn from its own input, eliminating the need for labeled training data. Extensive experiments demonstrate that the learned representations result in significant improvements when transferred to three different real-world tasks, particularly compared to task-specific modeling and multi-task representation learning