This paper reports on the NTIRE 2022 challenge on perceptual image quality assessment (IQA), held in conjunction with the New Trends in Image Restoration and Enhancement workshop (NTIRE) workshop at CVPR 2022. This challenge is held to address the emerging challenge of IQA by perceptual image processing algorithms. The output images of these algorithms have completely different characteristics from traditional distortions and are included in the PIPAL dataset used in this challenge. This challenge is divided into two tracks, a full-reference IQA track similar to the previous NTIRE IQA challenge and a new track that focuses on the no-reference IQA methods. The challenge has 192 and 179 registered participants for two tracks. In the final testing stage, 7 and 8 participating teams submitted their models and fact sheets. Almost all of them have achieved better results than existing IQA methods, and the winning method can demonstrate state-of-the-art performance.
Nine language-vision AI models trained on web scrapes with the Contrastive Language-Image Pretraining (CLIP) objective are evaluated for evidence of a bias studied by psychologists: the sexual objectification of girls and women, which occurs when a person's human characteristics are disregarded and the person is treated as a body or a collection of body parts. A first experiment uses standardized images of women from the Sexual OBjectification and EMotion Database, and finds that, commensurate with prior research in psychology, human characteristics are disassociated from images of objectified women: the model's recognition of emotional state is mediated by whether the subject is fully or partially clothed. Embedding association tests (EATs) return significant effect sizes for both anger (d >.8) and sadness (d >.5). A second experiment measures the effect in a representative application: an automatic image captioner (Antarctic Captions) includes words denoting emotion less than 50% as often for images of partially clothed women than for images of fully clothed women. A third experiment finds that images of female professionals (scientists, doctors, executives) are likely to be associated with sexual descriptions relative to images of male professionals. A fourth experiment shows that a prompt of "a [age] year old girl" generates sexualized images (as determined by an NSFW classifier) up to 73% of the time for VQGAN-CLIP (age 17), and up to 40% of the time for Stable Diffusion (ages 14 and 18); the corresponding rate for boys never surpasses 9%. The evidence indicates that language-vision AI models trained on automatically collected web scrapes learn biases of sexual objectification, which propagate to downstream applications.
The recent advances in diffusion models have set an impressive milestone in many generation tasks. Trending works such as DALL-E2, Imagen, and Stable Diffusion have attracted great interest in academia and industry. Despite the rapid landscape changes, recent new approaches focus on extensions and performance rather than capacity, thus requiring separate models for separate tasks. In this work, we expand the existing single-flow diffusion pipeline into a multi-flow network, dubbed Versatile Diffusion (VD), that handles text-to-image, image-to-text, image-variation, and text-variation in one unified model. Moreover, we generalize VD to a unified multi-flow multimodal diffusion framework with grouped layers, swappable streams, and other propositions that can process modalities beyond images and text. Through our experiments, we demonstrate that VD and its underlying framework have the following merits: a) VD handles all subtasks with competitive quality; b) VD initiates novel extensions and applications such as disentanglement of style and semantic, image-text dual-guided generation, etc.; c) Through these experiments and applications, VD provides more semantic insights of the generated outputs. Our code and models are open-sourced at https://github.com/SHI-Labs/Versatile-Diffusion.
In Vision-and-Language Navigation (VLN), researchers typically take an image encoder pre-trained on ImageNet without fine-tuning on the environments that the agent will be trained or tested on. However, the distribution shift between the training images from ImageNet and the views in the navigation environments may render the ImageNet pre-trained image encoder suboptimal. Therefore, in this paper, we design a set of structure-encoding auxiliary tasks (SEA) that leverage the data in the navigation environments to pre-train and improve the image encoder. Specifically, we design and customize (1) 3D jigsaw, (2) traversability prediction, and (3) instance classification to pre-train the image encoder. Through rigorous ablations, our SEA pre-trained features are shown to better encode structural information of the scenes, which ImageNet pre-trained features fail to properly encode but is crucial for the target navigation task. The SEA pre-trained features can be easily plugged into existing VLN agents without any tuning. For example, on Test-Unseen environments, the VLN agents combined with our SEA pre-trained features achieve absolute success rate improvement of 12% for Speaker-Follower, 5% for Env-Dropout, and 4% for AuxRN.
Medical image segmentation can be implemented using Deep Learning methods with fast and efficient segmentation networks. Single-board computers (SBCs) are difficult to use to train deep networks due to their memory and processing limitations. Specific hardware such as Google's Edge TPU makes them suitable for real time predictions using complex pre-trained networks. In this work, we study the performance of two SBCs, with and without hardware acceleration for fundus image segmentation, though the conclusions of this study can be applied to the segmentation by deep neural networks of other types of medical images. To test the benefits of hardware acceleration, we use networks and datasets from a previous published work and generalize them by testing with a dataset with ultrasound thyroid images. We measure prediction times in both SBCs and compare them with a cloud based TPU system. The results show the feasibility of Machine Learning accelerated SBCs for optic disc and cup segmentation obtaining times below 25 milliseconds per image using Edge TPUs.
Multimodal models trained on large natural image-text pair datasets have exhibited astounding abilities in generating high-quality images. Medical imaging data is fundamentally different to natural images, and the language used to succinctly capture relevant details in medical data uses a different, narrow but semantically rich, domain-specific vocabulary. Not surprisingly, multi-modal models trained on natural image-text pairs do not tend to generalize well to the medical domain. Developing generative imaging models faithfully representing medical concepts while providing compositional diversity could mitigate the existing paucity of high-quality, annotated medical imaging datasets. In this work, we develop a strategy to overcome the large natural-medical distributional shift by adapting a pre-trained latent diffusion model on a corpus of publicly available chest x-rays (CXR) and their corresponding radiology (text) reports. We investigate the model's ability to generate high-fidelity, diverse synthetic CXR conditioned on text prompts. We assess the model outputs quantitatively using image quality metrics, and evaluate image quality and text-image alignment by human domain experts. We present evidence that the resulting model (RoentGen) is able to create visually convincing, diverse synthetic CXR images, and that the output can be controlled to a new extent by using free-form text prompts including radiology-specific language. Fine-tuning this model on a fixed training set and using it as a data augmentation method, we measure a 5% improvement of a classifier trained jointly on synthetic and real images, and a 3% improvement when trained on a larger but purely synthetic training set. Finally, we observe that this fine-tuning distills in-domain knowledge in the text-encoder and can improve its representation capabilities of certain diseases like pneumothorax by 25%.
Recently, self-supervised learning (SSL) has achieved tremendous empirical advancements in learning image representation. However, our understanding and knowledge of the representation are still limited. This work shows that the success of the SOTA siamese-network-based SSL approaches is primarily based on learning a representation of image patches. Particularly, we show that when we learn a representation only for fixed-scale image patches and aggregate different patch representations linearly for an image (instance), it can achieve on par or even better results than the baseline methods on several benchmarks. Further, we show that the patch representation aggregation can also improve various SOTA baseline methods by a large margin. We also establish a formal connection between the SSL objective and the image patches co-occurrence statistics modeling, which supplements the prevailing invariance perspective. By visualizing the nearest neighbors of different image patches in the embedding space and projection space, we show that while the projection has more invariance, the embedding space tends to preserve more equivariance and locality. Finally, we propose a hypothesis for the future direction based on the discovery of this work.
Previous work has shown that a neural network with the rectified linear unit (ReLU) activation function leads to a convex polyhedral decomposition of the input space. These decompositions can be represented by a dual graph with vertices corresponding to polyhedra and edges corresponding to polyhedra sharing a facet, which is a subgraph of a Hamming graph. This paper illustrates how one can utilize the dual graph to detect and analyze adversarial attacks in the context of digital images. When an image passes through a network containing ReLU nodes, the firing or non-firing at a node can be encoded as a bit ($1$ for ReLU activation, $0$ for ReLU non-activation). The sequence of all bit activations identifies the image with a bit vector, which identifies it with a polyhedron in the decomposition and, in turn, identifies it with a vertex in the dual graph. We identify ReLU bits that are discriminators between non-adversarial and adversarial images and examine how well collections of these discriminators can ensemble vote to build an adversarial image detector. Specifically, we examine the similarities and differences of ReLU bit vectors for adversarial images, and their non-adversarial counterparts, using a pre-trained ResNet-50 architecture. While this paper focuses on adversarial digital images, ResNet-50 architecture, and the ReLU activation function, our methods extend to other network architectures, activation functions, and types of datasets.
Temporal volume images with 3D+t (4D) information are often used in medical imaging to statistically analyze temporal dynamics or capture disease progression. Although deep-learning-based generative models for natural images have been extensively studied, approaches for temporal medical image generation such as 4D cardiac volume data are limited. In this work, we present a novel deep learning model that generates intermediate temporal volumes between source and target volumes. Specifically, we propose a diffusion deformable model (DDM) by adapting the denoising diffusion probabilistic model that has recently been widely investigated for realistic image generation. Our proposed DDM is composed of the diffusion and the deformation modules so that DDM can learn spatial deformation information between the source and target volumes and provide a latent code for generating intermediate frames along a geodesic path. Once our model is trained, the latent code estimated from the diffusion module is simply interpolated and fed into the deformation module, which enables DDM to generate temporal frames along the continuous trajectory while preserving the topology of the source image. We demonstrate the proposed method with the 4D cardiac MR image generation between the diastolic and systolic phases for each subject. Compared to the existing deformation methods, our DDM achieves high performance on temporal volume generation.
Learning fine-grained interplay between vision and language allows to a more accurate understanding for VisionLanguage tasks. However, it remains challenging to extract key image regions according to the texts for semantic alignments. Most existing works are either limited by textagnostic and redundant regions obtained with the frozen detectors, or failing to scale further due to its heavy reliance on scarce grounding (gold) data to pre-train detectors. To solve these problems, we propose Self-Locator Aided Network (SLAN) for cross-modal understanding tasks without any extra gold data. SLAN consists of a region filter and a region adaptor to localize regions of interest conditioned on different texts. By aggregating cross-modal information, the region filter selects key regions and the region adaptor updates their coordinates with text guidance. With detailed region-word alignments, SLAN can be easily generalized to many downstream tasks. It achieves fairly competitive results on five cross-modal understanding tasks (e.g., 85.7% and 69.2% on COCO image-to-text and text-to-image retrieval, surpassing previous SOTA methods). SLAN also demonstrates strong zero-shot and fine-tuned transferability to two localization tasks.