JWJunfeng Wu

Junfeng Wu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Junfeng Wu has a total of 43 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Fish density estimation for smart marine aquaculture using a multi-granularity counting consistency network

    by Miao Chen, Junfeng Wu, Shaojiang Cheng, Saibei Zheng, Peihua Wu, Yida Sha, Yizhi Zhou, Miao Wang

    Fish density estimation is an important task in precision aquaculture, supporting feeding management, stocking regulation, and production monitoring. However, accurate estimation in aquaculture scenes remains challenging due to severe occlusion, lighting variation, water turbidity, and background interference. Existing fish density estimation methods mainly focus on image-level count accuracy or pixel-level density-map supervision but often lack explicit constraints on local count consistency, which may lead to inaccurate spatial distributions in dense scenes. Furthermore, features learned mainly from visual appearance are easily affected by complex visual backgrounds in aquaculture scenes. To address these issues, we propose MGCC-Net, a multi-granularity counting consistency network for fish density estimation in aquaculture scenes. MGCC-Net introduces two components: a multi-granularity local patch count loss to enforce local count consistency across multiple spatial scales and a dual-projection channel attention module that generates channel weights from two learned projections of the global visual descriptor and is jointly optimized with the density-estimation objective. We also construct AquaFish-Count, a real-world fish density dataset containing images from aquaculture and natural marine environments, organizing it into FD-High and FD-Low based on the annotated fish count in each image. Experimental results show that MGCC-Net achieves the best performance on FD-High and remains competitive on FD-Low, with its clearest advantage observed in dense scenes with severe fish overlap. These results demonstrate the effectiveness of MGCC-Net for fish density estimation and its potential application in marine biological environment perception and smart aquaculture monitoring systems.

  • Dynamic audio–visual fusion for accurate fish feeding intensity recognition in recirculating aquaculture environments

    by Yida Sha, Junfeng Wu, Saibei Zheng, Peihua Wu, Shaojiang Cheng, Miao Chen, Yizhi Zhou, Hong Yu


  • Overcoming body occlusion challenges: a novel underwater fish individual recognition based on local and global feature fusion

    by Xuelan Liang, Junfeng Wu, Yongzhi Xie, Miao Chen, Shaojiang Cheng, Yida Sha, Peihua Wu, Saibei Zheng, Hong Yu


  • Open Access

    Conceptual Validation of High-Precision Fish Feeding Behavior Recognition Using Semantic Segmentation and Real-Time Temporal Variance Analysis for Aquaculture

    by Han Kong, Junfeng Wu, Xuelan Liang, Yongzhi Xie, Boyu Qu, Hong Yu

    Aquaculture plays an important role in the global economy. However, unscientific feeding methods often lead to problems such as feed waste and water pollution. This study aims to address this issue by accurately recognizing fish feeding behaviors to provide automatic bait casting machines with scientific feeding strategies, thereby reducing farming costs. We propose a fish feeding behavior recognition method based on semantic segmentation, which overcomes the limitations of existing methods in dealing with complex backgrounds, water splash interference, fish target overlapping, and real-time performance. In this method, we first accurately segment fish targets in the images using a semantic segmentation model. Then, these segmented images are input into our proposed fish feeding behavior recognition model. By analyzing the aggregation characteristics during the feeding process, we can identify fish feeding behaviors. Experiments show that the proposed method has excellent robustness and real-time performance, and it performs well in the case of complex water background and occlusion of fish targets. We provide the aquaculture industry with an efficient and reliable method for recognizing fish feeding behavior, offering new scientific support for intelligent aquaculture and delivering powerful solutions to improve aquaculture management and production efficiency. Although the algorithm proposed in this study has shown good performance in fish feeding behavior recognition, it requires certain lighting conditions and fish density, which may affect its adaptability in different environments. Future research could explore integrating multimodal data, such as sound information, to assist in judgment, thereby enhancing the robustness of the model and promoting the development of intelligent aquaculture.


  • Fish‐Finder: A robust small target detection method for aquaculture fish in low‐quality underwater images

    by Liang Liu, Junfeng Wu, Haiyan Zhao, Han Kong, Tao Zheng, Boyu Qu, Hong Yu

    Underwater fish object detection serves as a pivotal research direction in marine biology, aquaculture management, and computer vision, yet it poses substantial challenges due to the complexity of underwater environments, occultations, and the small‐sized and frequently moving fish in aquaculture. Addressing these challenges, we propose a novel underwater fish object detection algorithm named Fish‐Finder. First, we engendered a structure titled “C2fBF,” utilizing the dual‐path routing attention protocol of BiFormer. The primary objective of this structure is to alleviate the perturbations induced by underwater intricacies during the phase of downsampling in the backbone network, thereby discerning and conserving finer contextual features. Subsequently, we co‐opted the RepGFPN method within our neck network—a distinctive approach that adeptly merges high‐level semantic constructs with low‐level spatial specifics, thus fortifying its multi‐scale detection prowess. Then, in an endeavor to diminish the sensitivity toward positional aberrations during the detection of diminutive aquatic creatures, we incorporated a novel bounding box regression loss function, the Wasserstein loss, to the existing CIoU. This innovative function gauges the congruity between the predicted bounding box Gaussian distribution and the reference bounding box Gaussian distribution. Finally, in regard to the dataset, we independently assembled a specific dataset termed “SmallFish.” This unique dataset, meticulously designed for the detection of small‐scale fish within intricate underwater settings, includes 5000 annotated images of small fish. Experimental results demonstrate that, compared to the state‐of‐the‐art detection methods, our proposed method improves the accuracy by and , and mean average precision (mAP) increases and in public dataset Kaggle‐Fish and our SmallFish dataset, respectively.


  • A novel fish individual recognition method for precision farming based on knowledge distillation strategy and the range of the receptive field

    by Jianhao Yin, Junfeng Wu, Chunqi Gao, Hong Yu, Liang Liu, Shihao Guo

    With the continuous development of green and high‐quality aquaculture technology, the process of industrialized aquaculture has been promoted. Automation, intelligence, and precision have become the future development trend of the aquaculture industry. Fish individual recognition can further distinguish fish individuals based on the determination of fish categories, providing basic support for fish disease analysis, bait feeding, and precision aquaculture. However, the high similarity of fish individuals and the complexity of the underwater environment presents great challenges to fish individual recognition. To address these problems, we propose a novel fish individual recognition method for precision farming that rethinks the knowledge distillation strategy and the chunking method in the vision transformer. The method uses the traditional convolutional neural network model as the teacher model, introducing the teacher token to guide the student model to learn the fish texture features. We propose stride patch embedding to expand the range of the receptive field, thus enhancing the local continuity of the image, and self‐attention‐pruning to discard unimportant tokens and reduce the model computation. The experimental results on the DlouFish dataset show that the proposed method in this paper improves accuracy by 3.25% compared to ECA Resnet152, with an accuracy of 93.19%, and also outperforms other vision transformer models.


  • Open Access

    A real-time feeding decision method based on density estimation of farmed fish

    by Haiyan Zhao, Junfeng Wu, Liang Liu, Boyu Qu, Jianhao Yin, Hong Yu, Zhongai Jiang, Chunyu Zhou

    With the global population growth and increasing demand for high-quality protein, aquaculture has experienced rapid development. Fish culture management and feed supply are crucial components of aquaculture. Traditional baiting management relies on experiential judgment and regular observation, which often leads to inefficient baiting practices and wastage. To address these issues, intelligent bait casting decisions have emerged. Leveraging advanced artificial intelligence algorithms, intelligent bait casting decisions can overcome most drawbacks of traditional bait management and enhance breeding efficiency. However, most of the current intelligent baiting decisions are focused on using methods such as image processing and target detection to identify different feeding actions and patterns. These methods do not discuss based on video streams and do not consider the changes in fish behavior during the baiting process. Therefore, we proposed a real-time analysis method based on the density distribution of fish feeding behavior (FishFeed). Firstly, this method upgrades the input mechanism, not only handling static images but also capable of real-time video stream analysis. Secondly, by evaluating the fish school density distribution through a new intelligent baiting strategy, this method can monitor the feeding behavior of fish school during the baiting process in real time. Finally, we constructed a dataset for fish school density analysis (DlouFishDensity) that includes a wealth of video and image frames, providing a valuable resource for research. Experimental results indicate that our algorithm outperforms MCNN, improving MAE by 1.63 and 1.35, MSE by 1.92 and 1.58, and reducing prediction time by 2.56 seconds on the same dataset. By implementing real-time analysis of fish feeding behavior density distribution, our method offers a more efficient and effective approach to baiting management in aquaculture, contributing to improved breeding efficiency and resource utilization.


  • Open Access

    Fish Recognition in the Underwater Environment Using an Improved ArcFace Loss for Precision Aquaculture

    by Liang Liu, Junfeng Wu, Tao Zheng, Haiyan Zhao, Han Kong, Boyu Qu, Hong Yu

    Accurate fish individual recognition is one of the critical technologies for large-scale fishery farming when trying to achieve accurate, green farming and sustainable development. It is an essential link for aquaculture to move toward automation and intelligence. However, existing fish individual data collection methods cannot cope with the interference of light, blur, and pose in the natural underwater environment, which makes the captured fish individual images of poor quality. These low-quality images can cause significant interference with the training of recognition networks. In order to solve the above problems, this paper proposes an underwater fish individual recognition method (FishFace) that combines data quality assessment and loss weighting. First, we introduce the Gem pooing and quality evaluation module, which is based on EfficientNet. This module is an improved fish recognition network that can evaluate the quality of fish images well, and it does not need additional labels; second, we propose a new loss function, FishFace Loss, which will weigh the loss according to the quality of the image so that the model focuses more on recognizable fish images, and less on images that are difficult to recognize. Finally, we collect a dataset for fish individual recognition (WideFish), which contains and annotates 5000 images of 300 fish. The experimental results show that, compared with the state-of-the-art individual recognition methods, Rank1 accuracy is improved by 2.60% and 3.12% on the public dataset DlouFish and the proposed WideFish dataset, respectively.


  • Open Access

    Individual Fish Recognition Method with Coarse and Fine-Grained Feature Linkage Learning for Precision Aquaculture

    by Jianhao Yin, Junfeng Wu, Chunqi Gao, Hong Yu, Liang Liu, Zhongai Jiang, Shihao Guo

    With the increasing level of precision and intelligence in the aquaculture, real-time mastery of the growth status of aquaculture individuals has become an important means to improve aquaculture efficiency and save resources and the environment. Therefore, accurate individual recognition of underwater fish has become one of the key technologies for precision aquaculture. In order to cope with the impact of the complex underwater environment on the recognition accuracy, this paper proposes a coarse and fine-grained features learning method for individual fish recognition. The method consists of a coarse-grained feature learning network and two fine-grained feature learning networks. The trunk of the network is responsible for learning coarse-grained features of the fish, the first branch learns fine-grained features of fish from head, body, and tail, and the second branch learns fine-grained features of fish from upper and lower fins. we supplemented different levels of noise and attack to the training set of fine-grained features and enriched the grayscale variation to cope with the complexity and variability of the underwater environment. The simulation experimental results show that the method achieves more than 96.7% in key indicators such as Rank-1 and Rank-5, and also performs well in other fish recognition tasks with certain generalization.