The development and adoption of Vision Transformers and other deep-learning architectures for image classification tasks has been rapid. However, the "black box" nature of neural networks is a barrier to adoption in applications where explainability is essential. While some techniques for generating explanations have been proposed, primarily for Convolutional Neural Networks, adapting such techniques to the new paradigm of Vision Transformers is non-trivial. This paper presents T-TAME, Transformer-compatible Trainable Attention Mechanism for Explanations, a general methodology for explaining deep neural networks used in image classification tasks. The proposed architecture and training technique can be easily applied to any convolutional or Vision Transformer-like neural network, using a streamlined training approach. After training, explanation maps can be computed in a single forward pass; these explanation maps are comparable to or outperform the outputs of computationally expensive perturbation-based explainability techniques, achieving SOTA performance. We apply T-TAME to three popular deep learning classifier architectures, VGG-16, ResNet-50, and ViT-B-16, trained on the ImageNet dataset, and we demonstrate improvements over existing state-of-the-art explainability methods. A detailed analysis of the results and an ablation study provide insights into how the T-TAME design choices affect the quality of the generated explanation maps.
This paper presents a web-based tool that facilitates the production of tailored summaries for online sharing on social media. Through an interactive user interface, it supports a ``one-click'' video summarization process. Based on the integrated AI models for video summarization and aspect ratio transformation, it facilitates the generation of multiple summaries of a full-length video according to the needs of target platforms with regard to the video's length and aspect ratio.
In this work, we present an integrated system for spatiotemporal summarization of 360-degrees videos. The video summary production mainly involves the detection of salient events and their synopsis into a concise summary. The analysis relies on state-of-the-art methods for saliency detection in 360-degrees video (ATSal and SST-Sal) and video summarization (CA-SUM). It also contains a mechanism that classifies a 360-degrees video based on the use of static or moving camera during recording and decides which saliency detection method will be used, as well as a 2D video production component that is responsible to create a conventional 2D video containing the salient events in the 360-degrees video. Quantitative evaluations using two datasets for 360-degrees video saliency detection (VR-EyeTracking, Sports-360) show the accuracy and positive impact of the developed decision mechanism, and justify our choice to use two different methods for detecting the salient events. A qualitative analysis using content from these datasets, gives further insights about the functionality of the decision mechanism, shows the pros and cons of each used saliency detection method and demonstrates the advanced performance of the trained summarization method against a more conventional approach.
Recent image manipulation localization and detection techniques usually leverage forensic artifacts and traces that are produced by a noise-sensitive filter, such as SRM and Bayar convolution. In this paper, we showcase that different filters commonly used in such approaches excel at unveiling different types of manipulations and provide complementary forensic traces. Thus, we explore ways of merging the outputs of such filters and aim to leverage the complementary nature of the artifacts produced to perform image manipulation localization and detection (IMLD). We propose two distinct methods: one that produces independent features from each forensic filter and then fuses them (this is referred to as late fusion) and one that performs early mixing of different modal outputs and produces early combined features (this is referred to as early fusion). We demonstrate that both approaches achieve competitive performance for both image manipulation localization and detection, outperforming state-of-the-art models across several datasets.
Face detectors are becoming a crucial component of many applications, including surveillance, that often have to run on edge devices with limited processing power and memory. Therefore, there's a pressing demand for compact face detection models that can function efficiently across resource-constrained devices. Over recent years, network pruning techniques have attracted a lot of attention from researchers. These methods haven't been well examined in the context of face detectors, despite their expanding popularity. In this paper, we implement filter pruning on two already small and compact face detectors, named EXTD (Extremely Tiny Face Detector) and EResFD (Efficient ResNet Face Detector). The main pruning algorithm that we utilize is Filter Pruning via Geometric Median (FPGM), combined with the Soft Filter Pruning (SFP) iterative procedure. We also apply L1 Norm pruning, as a baseline to compare with the proposed approach. The experimental evaluation on the WIDER FACE dataset indicates that the proposed approach has the potential to further reduce the model size of already lightweight face detectors, with limited accuracy loss, or even with small accuracy gain for low pruning rates.
In this paper, we introduce Masked Feature Modelling (MFM), a novel approach for the unsupervised pre-training of a Graph Attention Network (GAT) block. MFM utilizes a pretrained Visual Tokenizer to reconstruct masked features of objects within a video, leveraging the MiniKinetics dataset. We then incorporate the pre-trained GAT block into a state-of-the-art bottom-up supervised video-event recognition architecture, ViGAT, to improve the model's starting point and overall accuracy. Experimental evaluations on the YLI-MED dataset demonstrate the effectiveness of MFM in improving event recognition performance.
In this paper, Gated-ViGAT, an efficient approach for video event recognition, utilizing bottom-up (object) information, a new frame sampling policy and a gating mechanism is proposed. Specifically, the frame sampling policy uses weighted in-degrees (WiDs), derived from the adjacency matrices of graph attention networks (GATs), and a dissimilarity measure to select the most salient and at the same time diverse frames representing the event in the video. Additionally, the proposed gating mechanism fetches the selected frames sequentially, and commits early-exiting when an adequately confident decision is achieved. In this way, only a few frames are processed by the computationally expensive branch of our network that is responsible for the bottom-up information extraction. The experimental evaluation on two large, publicly available video datasets (MiniKinetics, ActivityNet) demonstrates that Gated-ViGAT provides a large computational complexity reduction in comparison to our previous approach (ViGAT), while maintaining the excellent event recognition and explainability performance. Gated-ViGAT source code is made publicly available at https://github.com/bmezaris/Gated-ViGAT
The apparent ``black box'' nature of neural networks is a barrier to adoption in applications where explainability is essential. This paper presents TAME (Trainable Attention Mechanism for Explanations), a method for generating explanation maps with a multi-branch hierarchical attention mechanism. TAME combines a target model's feature maps from multiple layers using an attention mechanism, transforming them into an explanation map. TAME can easily be applied to any convolutional neural network (CNN) by streamlining the optimization of the attention mechanism's training method and the selection of target model's feature maps. After training, explanation maps can be computed in a single forward pass. We apply TAME to two widely used models, i.e. VGG-16 and ResNet-50, trained on ImageNet and show improvements over previous top-performing methods. We also provide a comprehensive ablation study comparing the performance of different variations of TAME's architecture. TAME source code is made publicly available at https://github.com/bmezaris/TAME
In this paper we tackle the cross-modal video retrieval problem and, more specifically, we focus on text-to-video retrieval. We investigate how to optimally combine multiple diverse textual and visual features into feature pairs that lead to generating multiple joint feature spaces, which encode text-video pairs into comparable representations. To learn these representations our proposed network architecture is trained by following a multiple space learning procedure. Moreover, at the retrieval stage, we introduce additional softmax operations for revising the inferred query-video similarities. Extensive experiments in several setups based on three large-scale datasets (IACC.3, V3C1, and MSR-VTT) lead to conclusions on how to best combine text-visual features and document the performance of the proposed network. Source code is made publicly available at: https://github.com/bmezaris/TextToVideoRetrieval-TtimesV
In this paper two new learning-based eXplainable AI (XAI) methods for deep convolutional neural network (DCNN) image classifiers, called L-CAM-Fm and L-CAM-Img, are proposed. Both methods use an attention mechanism that is inserted in the original (frozen) DCNN and is trained to derive class activation maps (CAMs) from the last convolutional layer's feature maps. During training, CAMs are applied to the feature maps (L-CAM-Fm) or the input image (L-CAM-Img) forcing the attention mechanism to learn the image regions explaining the DCNN's outcome. Experimental evaluation on ImageNet shows that the proposed methods achieve competitive results while requiring a single forward pass at the inference stage. Moreover, based on the derived explanations a comprehensive qualitative analysis is performed providing valuable insight for understanding the reasons behind classification errors, including possible dataset biases affecting the trained classifier.