Most image retrieval works aim at learning discriminative visual features, while little attention is paid to the retrieval efficiency. The speed of feature extraction is key to the real-world system. Therefore, in this paper, we focus on network pruning for image retrieval acceleration. Different from the classification models predicting discrete categories, image retrieval models usually extract continuous features for retrieval, which are more sensitive to network pruning. Such different characteristics of the retrieval and classification models make the traditional pruning method sub-optimal for image retrieval acceleration. Two points are critical for pruning image retrieval models: preserving the local geometry structure of filters and maintaining the model capacity during pruning. In view of the above considerations, we propose a Progressive Local Filter Pruning (PLFP) method. Specifically, we analyze the local geometry of filter distribution in every layer and select redundant filters according to one new criterion that the filter can be replaced locally by other similar filters. Furthermore, to preserve the model capacity of the original model, the proposed method progressively prune the filter by decreasing the scale of filter weights gradually. We evaluate our method on four scene retrieval datasets, i.e. , Oxford5K, Oxford105 K, Paris6K, and Paris106 K, and one person re-identification dataset, i.e. , Market-1501. Extensive experiments show that the proposed method (1) preserves the original model capacity while pruning (2) and achieves superior performance to other widely-used pruning methods.

Video-based pedestrian re-identification is a fundamental task in video surveillance and real-world applications, and has attracted much research attention recently. Its goal is to match pedestrians across multiple non-overlapping network cameras. In this paper we propose a novel two-stream multi-rate recurrent neural network for video-based pedestrian re-identification, which has two inherent benefits: (1) capture the static spatial and temporal information; (2) deal with motion speed variance. Given video sequences of pedestrians, we start with extracting spatial and motion features using two different deep neural networks. Then we combine them using a regularized fusion network, which aims to explore feature correlations. To step further, we feed the two features into a multi-rate recurrent network to exploit the temporal correlations, and more importantly, to take into consideration that pedestrians, sometimes even the same pedestrian, move in different speeds across different camera views. Extensive experiments have conducted on two real-world video-based pedestrian re-identification benchmarks: iLIDS-VID and PRID 2011 datasets. The experimental results confirm the superiority of the proposed method. Our code will be released upon acceptance.

Place recognition is one of the most fundamental topics in the computer-vision and robotics communities, where the task is to accurately and efficiently recognize the location of a given query image. Despite years of knowledge accumulated in this field, place recognition still remains an open problem due to the various ways in which the appearance of real-world places may differ. This paper presents an overview of the place-recognition literature. Since condition-invariant and viewpoint-invariant features are essential factors to long-term robust visual place-recognition systems, we start with traditional image-description methodology developed in the past, which exploits techniques from the image-retrieval field. Recently, the rapid advances of related fields, such as object detection and image classification, have inspired a new technique to improve visual place-recognition systems, that is, convolutional neural networks (CNNs). Thus, we then introduce the recent progress of visual place-recognition systems based on CNNs to automatically learn better image representations for places. Finally, we close with discussions and mention of future work on place recognition.

Industrial visual inspection plays a crucial role in intelligent manufacturing. However, existing anomaly-detection methods based on unsupervised learning paradigms often struggle with issues such as overlooking minor defects and blurring component edges in confidence maps. To address these challenges, this paper proposes an industrial anomaly-detection method based on component-level feature enhancement. This method introduces a component-level feature-enhancement module, which optimizes feature matching by calculating the structural similarity between global coarse-grained confidence features and local fine-grained confidence features, thereby generating enhanced feature maps to improve the model’s detection accuracy for minor defects and local anomalies. Additionally, we propose a region-segmentation method based on multi-layer piecewise thresholds, which effectively distinguishes between foreground and background in confidence maps, circumvents background interference and ensures the integrity of structural information of foreground components. Experimental results demonstrate that the proposed method surpasses comparative methods in both logical and structural defect detection tasks, showing significant advantages, especially in fine-grained anomaly detection, with stronger robustness and accuracy.