Search

In computer vision, considerable success has been achieved in image classification. Yet, real-world applications can be negatively affected by noise corruption, so a method for dealing with noise is crucial. Knowledge distillation that utilizes the knowledge of a teacher model trained on clean images to train a student model on noisy images is a promising technique because it can be applied without special modification of the classifier.

Deep neural networks present their potential vulnerabilities to backdoor attacks. They have a satisfactory performance for benign users with clean samples but will get malicious outputs when inputs are attached with the backdoor trigger. Current backdoor attacks on image classifiers usually target only one single class, making them not robust to defenses against this characteristic.

Multi-task learning (MTL) permits joint task learning based on a shared deep learning architecture and multiple loss functions. Despite the recent advances in MTL, one loss often dominates the learning optimization in multiple unrelated tasks. This often results in poor performance compared to the corresponding single task learning.

In this paper, we introduce 3D-CSL, a compact pipeline for Near-Duplicate Video Retrieval (NDVR), and explore a novel self-supervised learning strategy for video similarity learning. Most previous NDVR methods depend a lot on pair-wise labeled data, so that be limited by the scale of datasets and cannot optimize complex but efficient backbones, e.g., 3D transformers.

Video object detection is a challenging task due to deteriorated object appearances. To enhance per-frame features, one way is to aggregate features from several support frames. However, proposals generated by the region proposal network may not be precise and diverse due to the fixed anchors, limiting the detection performance.

This paper proposes InteractionLIME: a model-agnostic attribution technique to explain deep models predictions in terms of feature interactions. Specifically, we regress a bilinear form to approximate the output of any two-input model, by sampling pertubations of both inputs simultaneously. Upon training, we retrieve a global explanation and a set of feature partitioning maps via the singular value decomposition of the learned interaction matrix of the bilinear model.

Deep learning models trained on long-tailed (LT) datasets often exhibit bias towards head classes with high frequency. This paper highlights the limitations of existing solutions that combine class- and instance-level re-weighting loss in a naive manner. Specifically, we demonstrate that such solutions result in overfitting the training set, significantly impacting the rare classes.

In Siamese-network-based self-supervised learning (SSL), multi-level supervision (MLS) is a natural extension to enforce intermediate representations’ consistency against data augmentations. Although existing studies incorporated MLS to boost their system performances in combination with other ideas, simple MLS without bells and whistles is not deeply analyzed.

Infrared (IR) small target detection is a challenging task due to the low contrast and low signal-to-noise ratio, generally yields high false alarm rates. To improve the performance of IR small target detection, we propose a saliency guided multi-task leaning model (SGMTLM). The model consists of two parts: feature fusion and saliency detection. The feature fusion module is to integrate shallow information and deep semantic information of small targets.

Ship detection in synthetic aperture radar (SAR) images is a major issue in naval surveillance and port management. Existing challenges are mainly as follows: (1) Tiny ships are mixed with scattered noise spots on the sea. (2) Ships come in extreme aspect-ratios and various scales. (3) The land background blurs the outline of coastal ships.