Self-supervised learning methods are gaining increasing traction in computer vision due to their recent success in reducing the gap with supervised learning. In natural language processing (NLP) self-supervised learning and transformers are already t
he methods of choice. The recent literature suggests that the transformers are becoming increasingly popular also in computer vision. So far, the vision transformers have been shown to work well when pretrained either using a large scale supervised data or with some kind of co-supervision, e.g. in terms of teacher network. These supervised pretrained vision transformers achieve very good results in downstream tasks with minimal changes. In this work we investigate the merits of self-supervised learning for pretraining image/vision transformers and then using them for downstream classification tasks. We propose Self-supervised vIsion Transformers (SiT) and discuss several self-supervised training mechanisms to obtain a pretext model. The architectural flexibility of SiT allows us to use it as an autoencoder and work with multiple self-supervised tasks seamlessly. We show that a pretrained SiT can be finetuned for a downstream classification task on small scale datasets, consisting of a few thousand images rather than several millions. The proposed approach is evaluated on standard datasets using common protocols. The results demonstrate the strength of the transformers and their suitability for self-supervised learning. We outperformed existing self-supervised learning methods by large margin. We also observed that SiT is good for few shot learning and also showed that it is learning useful representation by simply training a linear classifier on top of the learned features from SiT. Pretraining, finetuning, and evaluation codes will be available under: https://github.com/Sara-Ahmed/SiT.
In the image fusion field, the design of deep learning-based fusion methods is far from routine. It is invariably fusion-task specific and requires a careful consideration. The most difficult part of the design is to choose an appropriate strategy to
generate the fused image for a specific task in hand. Thus, devising learnable fusion strategy is a very challenging problem in the community of image fusion. To address this problem, a novel end-to-end fusion network architecture (RFN-Nest) is developed for infrared and visible image fusion. We propose a residual fusion network (RFN) which is based on a residual architecture to replace the traditional fusion approach. A novel detail-preserving loss function, and a feature enhancing loss function are proposed to train RFN. The fusion model learning is accomplished by a novel two-stage training strategy. In the first stage, we train an auto-encoder based on an innovative nest connection (Nest) concept. Next, the RFN is trained using the proposed loss functions. The experimental results on public domain data sets show that, compared with the existing methods, our end-to-end fusion network delivers a better performance than the state-of-the-art methods in both subjective and objective evaluation. The code of our fusion method is available at https://github.com/hli1221/imagefusion-rfn-nest
Automated neural network design has received ever-increasing attention with the evolution of deep convolutional neural networks (CNNs), especially involving their deployment on embedded and mobile platforms. One of the biggest problems that neural ar
chitecture search (NAS) confronts is that a large number of candidate neural architectures are required to train, using, for instance, reinforcement learning and evolutionary optimisation algorithms, at a vast computation cost. Even recent differentiable neural architecture search (DNAS) samples a small number of candidate neural architectures based on the probability distribution of learned architecture parameters to select the final neural architecture. To address this computational complexity issue, we introduce a novel emph{architecture parameterisation} based on scaled sigmoid function, and propose a general emph{Differentiable Neural Architecture Learning} (DNAL) method to optimize the neural architecture without the need to evaluate candidate neural networks. Specifically, for stochastic supernets as well as conventional CNNs, we build a new channel-wise module layer with the architecture components controlled by a scaled sigmoid function. We train these neural network models from scratch. The network optimization is decoupled into the weight optimization and the architecture optimization. We address the non-convex optimization problem of neural architecture by the continuous scaled sigmoid method with convergence guarantees. Extensive experiments demonstrate our DNAL method delivers superior performance in terms of neural architecture search cost. The optimal networks learned by DNAL surpass those produced by the state-of-the-art methods on the benchmark CIFAR-10 and ImageNet-1K dataset in accuracy, model size and computational complexity.
Image decomposition is a crucial subject in the field of image processing. It can extract salient features from the source image. We propose a new image decomposition method based on convolutional neural network. This method can be applied to many im
age processing tasks. In this paper, we apply the image decomposition network to the image fusion task. We input infrared image and visible light image and decompose them into three high-frequency feature images and a low-frequency feature image respectively. The two sets of feature images are fused using a specific fusion strategy to obtain fusion feature images. Finally, the feature images are reconstructed to obtain the fused image. Compared with the state-of-the-art fusion methods, this method has achieved better performance in both subjective and objective evaluation.
Cross-modal person re-identification (Re-ID) is critical for modern video surveillance systems. The key challenge is to align inter-modality representations according to semantic information present for a person and ignore background information. In
this work, we present AXM-Net, a novel CNN based architecture designed for learning semantically aligned visual and textual representations. The underlying building block consists of multiple streams of feature maps coming from visual and textual modalities and a novel learnable context sharing semantic alignment network. We also propose complementary intra modal attention learning mechanisms to focus on more fine-grained local details in the features along with a cross-modal affinity loss for robust feature matching. Our design is unique in its ability to implicitly learn feature alignments from data. The entire AXM-Net can be trained in an end-to-end manner. We report results on both person search and cross-modal Re-ID tasks. Extensive experimentation validates the proposed framework and demonstrates its superiority by outperforming the current state-of-the-art methods by a significant margin.
Although group convolution operators are increasingly used in deep convolutional neural networks to improve the computational efficiency and to reduce the number of parameters, most existing methods construct their group convolution architectures by
a predefined partitioning of the filters of each convolutional layer into multiple regular filter groups with an equal spatial group size and data-independence, which prevents a full exploitation of their potential. To tackle this issue, we propose a novel method of designing self-grouping convolutional neural networks, called SG-CNN, in which the filters of each convolutional layer group themselves based on the similarity of their importance vectors. Concretely, for each filter, we first evaluate the importance value of their input channels to identify the importance vectors, and then group these vectors by clustering. Using the resulting emph{data-dependent} centroids, we prune the less important connections, which implicitly minimizes the accuracy loss of the pruning, thus yielding a set of emph{diverse} group convolution filters. Subsequently, we develop two fine-tuning schemes, i.e. (1) both local and global fine-tuning and (2) global only fine-tuning, which experimentally deliver comparable results, to recover the recognition capacity of the pruned network. Comprehensive experiments carried out on the CIFAR-10/100 and ImageNet datasets demonstrate that our self-grouping convolution method adapts to various state-of-the-art CNN architectures, such as ResNet and DenseNet, and delivers superior performance in terms of compression ratio, speedup and recognition accuracy. We demonstrate the ability of SG-CNN to generalise by transfer learning, including domain adaption and object detection, showing competitive results. Our source code is available at https://github.com/QingbeiGuo/SG-CNN.git.
We propose a novel end-to-end semi-supervised adversarial framework to generate photorealistic face images of new identities with wide ranges of expressions, poses, and illuminations conditioned by a 3D morphable model. Previous adversarial style-tra
nsfer methods either supervise their networks with large volume of paired data or use unpaired data with a highly under-constrained two-way generative framework in an unsupervised fashion. We introduce pairwise adversarial supervision to constrain two-way domain adaptation by a small number of paired real and synthetic images for training along with the large volume of unpaired data. Extensive qualitative and quantitative experiments are performed to validate our idea. Generated face images of new identities contain pose, lighting and expression diversity and qualitative results show that they are highly constraint by the synthetic input image while adding photorealism and retaining identity information. We combine face images generated by the proposed method with the real data set to train face recognition algorithms. We evaluated the model on two challenging data sets: LFW and IJB-A. We observe that the generated images from our framework consistently improves over the performance of deep face recognition network trained with Oxford VGG Face dataset and achieves comparable results to the state-of-the-art.
This paper investigates the evaluation of dense 3D face reconstruction from a single 2D image in the wild. To this end, we organise a competition that provides a new benchmark dataset that contains 2000 2D facial images of 135 subjects as well as the
ir 3D ground truth face scans. In contrast to previous competitions or challenges, the aim of this new benchmark dataset is to evaluate the accuracy of a 3D dense face reconstruction algorithm using real, accurate and high-resolution 3D ground truth face scans. In addition to the dataset, we provide a standard protocol as well as a Python script for the evaluation. Last, we report the results obtained by three state-of-the-art 3D face reconstruction systems on the new benchmark dataset. The competition is organised along with the 2018 13th IEEE Conference on Automatic Face & Gesture Recognition.
We present a new loss function, namely Wing loss, for robust facial landmark localisation with Convolutional Neural Networks (CNNs). We first compare and analyse different loss functions including L2, L1 and smooth L1. The analysis of these loss func
tions suggests that, for the training of a CNN-based localisation model, more attention should be paid to small and medium range errors. To this end, we design a piece-wise loss function. The new loss amplifies the impact of errors from the interval (-w, w) by switching from L1 loss to a modified logarithm function. To address the problem of under-representation of samples with large out-of-plane head rotations in the training set, we propose a simple but effective boosting strategy, referred to as pose-based data balancing. In particular, we deal with the data imbalance problem by duplicating the minority training samples and perturbing them by injecting random image rotation, bounding box translation and other data augmentation approaches. Last, the proposed approach is extended to create a two-stage framework for robust facial landmark localisation. The experimental results obtained on AFLW and 300W demonstrate the merits of the Wing loss function, and prove the superiority of the proposed method over the state-of-the-art approaches.
We present a framework for robust face detection and landmark localisation of faces in the wild, which has been evaluated as part of `the 2nd Facial Landmark Localisation Competition. The framework has four stages: face detection, bounding box aggreg
ation, pose estimation and landmark localisation. To achieve a high detection rate, we use two publicly available CNN-based face detectors and two proprietary detectors. We aggregate the detected face bounding boxes of each input image to reduce false positives and improve face detection accuracy. A cascaded shape regressor, trained using faces with a variety of pose variations, is then employed for pose estimation and image pre-processing. Last, we train the final cascaded shape regressor for fine-grained landmark localisation, using a large number of training samples with limited pose variations. The experimental results obtained on the 300W and Menpo benchmarks demonstrate the superiority of our framework over state-of-the-art methods.
