Noise reduction is one the most important and still active research topic in low-level image processing due to its high impact on object detection and scene understanding for computer vision systems. Recently, we can observe a substantial increase of interest in the application of deep learning algorithms in many computer vision problems due to its impressive capability of automatic feature extraction and classification. These methods have been also successfully applied in image denoising, significantly improving the performance, but most of the proposed approaches were designed for Gaussian noise suppression. In this paper, we present a switching filtering design intended for impulsive noise removal using deep learning. In the proposed method, the impulses are identified using a novel deep neural network architecture and noisy pixels are restored using the fast adaptive mean filter. The performed experiments show that the proposed approach is superior to the state-of-the-art filters designed for impulsive noise removal in digital color images.
In a typical multi-label setting, a picture contains on average few positive labels, and many negative ones. This positive-negative imbalance dominates the optimization process, and can lead to under-emphasizing gradients from positive labels during training, resulting in poor accuracy. In this paper, we introduce a novel asymmetric loss ("ASL"), which operates differently on positive and negative samples. The loss enables to dynamically down-weights and hard-thresholds easy negative samples, while also discarding possibly mislabeled samples. We demonstrate how ASL can balance the probabilities of different samples, and how this balancing is translated to better mAP scores. With ASL, we reach state-of-the-art results on multiple popular multi-label datasets: MS-COCO, Pascal-VOC, NUS-WIDE and Open Images. We also demonstrate ASL applicability for other tasks, such as single-label classification and object detection. ASL is effective, easy to implement, and does not increase the training time or complexity. Implementation is available at: https://github.com/Alibaba-MIIL/ASL.
While self-supervised learning has been shown to benefit a number of vision tasks, existing techniques mainly focus on image-level manipulation, which may not generalize well to downstream tasks at patch or pixel levels. Moreover, existing SSL methods might not sufficiently describe and associate the above representations within and across image scales. In this paper, we propose a Self-Supervised Pyramid Representation Learning (SS-PRL) framework. The proposed SS-PRL is designed to derive pyramid representations at patch levels via learning proper prototypes, with additional learners to observe and relate inherent semantic information within an image. In particular, we present a cross-scale patch-level correlation learning in SS-PRL, which allows the model to aggregate and associate information learned across patch scales. We show that, with our proposed SS-PRL for model pre-training, one can easily adapt and fine-tune the models for a variety of applications including multi-label classification, object detection, and instance segmentation.
Segment Anything Model 2 (SAM2) demonstrates exceptional performance in video segmentation and refinement of segmentation results. We anticipate that it can further evolve to achieve higher levels of automation for practical applications. Building upon SAM2, we conducted a series of practices that ultimately led to the development of a fully automated pipeline, termed Det-SAM2, in which object prompts are automatically generated by a detection model to facilitate inference and refinement by SAM2. This pipeline enables inference on infinitely long video streams with constant VRAM and RAM usage, all while preserving the same efficiency and accuracy as the original SAM2. This technical report focuses on the construction of the overall Det-SAM2 framework and the subsequent engineering optimization applied to SAM2. We present a case demonstrating an application built on the Det-SAM2 framework: AI refereeing in a billiards scenario, derived from our business context. The project at \url{https://github.com/motern88/Det-SAM2}.
In this paper, we propose to exploit the rich hierarchical features of deep convolutional neural networks to improve the accuracy and robustness of visual tracking. Deep neural networks trained on object recognition datasets consist of multiple convolutional layers. These layers encode target appearance with different levels of abstraction. For example, the outputs of the last convolutional layers encode the semantic information of targets and such representations are invariant to significant appearance variations. However, their spatial resolutions are too coarse to precisely localize the target. In contrast, features from earlier convolutional layers provide more precise localization but are less invariant to appearance changes. We interpret the hierarchical features of convolutional layers as a nonlinear counterpart of an image pyramid representation and explicitly exploit these multiple levels of abstraction to represent target objects. Specifically, we learn adaptive correlation filters on the outputs from each convolutional layer to encode the target appearance. We infer the maximum response of each layer to locate targets in a coarse-to-fine manner. To further handle the issues with scale estimation and re-detecting target objects from tracking failures caused by heavy occlusion or out-of-the-view movement, we conservatively learn another correlation filter, that maintains a long-term memory of target appearance, as a discriminative classifier. We apply the classifier to two types of object proposals: (1) proposals with a small step size and tightly around the estimated location for scale estimation; and (2) proposals with large step size and across the whole image for target re-detection. Extensive experimental results on large-scale benchmark datasets show that the proposed algorithm performs favorably against state-of-the-art tracking methods.
Driven by the rapid development of deep learning technology, the YOLO series has set a new benchmark for real-time object detectors. Additionally, transformer-based structures have emerged as the most powerful solution in the field, greatly extending the model's receptive field and achieving significant performance improvements. However, this improvement comes at a cost as the quadratic complexity of the self-attentive mechanism increases the computational burden of the model. To address this problem, we introduce a simple yet effective baseline approach called Mamba YOLO. Our contributions are as follows: 1) We propose that the ODMamba backbone introduce a \textbf{S}tate \textbf{S}pace \textbf{M}odel (\textbf{SSM}) with linear complexity to address the quadratic complexity of self-attention. Unlike the other Transformer-base and SSM-base method, ODMamba is simple to train without pretraining. 2) For real-time requirement, we designed the macro structure of ODMamba, determined the optimal stage ratio and scaling size. 3) We design the RG Block that employs a multi-branch structure to model the channel dimensions, which addresses the possible limitations of SSM in sequence modeling, such as insufficient receptive fields and weak image localization. This design captures localized image dependencies more accurately and significantly. Extensive experiments on the publicly available COCO benchmark dataset show that Mamba YOLO achieves state-of-the-art performance compared to previous methods. Specifically, a tiny version of Mamba YOLO achieves a \textbf{7.5}\% improvement in mAP on a single 4090 GPU with an inference time of \textbf{1.5} ms. The pytorch code is available at: \url{https://github.com/HZAI-ZJNU/Mamba-YOLO}
Deep learning has largely reshaped remote sensing (RS) research for aerial image understanding and made a great success. Nevertheless, most of the existing deep models are initialized with the ImageNet pretrained weights. Since natural images inevitably present a large domain gap relative to aerial images, probably limiting the finetuning performance on downstream aerial scene tasks. This issue motivates us to conduct an empirical study of remote sensing pretraining (RSP) on aerial images. To this end, we train different networks from scratch with the help of the largest RS scene recognition dataset up to now -- MillionAID, to obtain a series of RS pretrained backbones, including both convolutional neural networks (CNN) and vision transformers such as Swin and ViTAE, which have shown promising performance on computer vision tasks. Then, we investigate the impact of RSP on representative downstream tasks including scene recognition, semantic segmentation, object detection, and change detection using these CNN and vision transformer backbones. Empirical study shows that RSP can help deliver distinctive performances in scene recognition tasks and in perceiving RS related semantics such as "Bridge" and "Airplane". We also find that, although RSP mitigates the data discrepancies of traditional ImageNet pretraining on RS images, it may still suffer from task discrepancies, where downstream tasks require different representations from scene recognition tasks. These findings call for further research efforts on both large-scale pretraining datasets and effective pretraining methods. The codes and pretrained models will be released at https://github.com/ViTAE-Transformer/ViTAE-Transformer-Remote-Sensing.
Automatic captioning of images is a task that combines the challenges of image analysis and text generation. One important aspect in captioning is the notion of attention: How to decide what to describe and in which order. Inspired by the successes in text analysis and translation, previous work have proposed the \textit{transformer} architecture for image captioning. However, the structure between the \textit{semantic units} in images (usually the detected regions from object detection model) and sentences (each single word) is different. Limited work has been done to adapt the transformer's internal architecture to images. In this work, we introduce the \textbf{\textit{image transformer}}, which consists of a modified encoding transformer and an implicit decoding transformer, motivated by the relative spatial relationship between image regions. Our design widen the original transformer layer's inner architecture to adapt to the structure of images. With only regions feature as inputs, our model achieves new state-of-the-art performance on both MSCOCO offline and online testing benchmarks.
The non-local block is a popular module for strengthening the context modeling ability of a regular convolutional neural network. This paper first studies the non-local block in depth, where we find that its attention computation can be split into two terms, a whitened pairwise term accounting for the relationship between two pixels and a unary term representing the saliency of every pixel. We also observe that the two terms trained alone tend to model different visual clues, e.g. the whitened pairwise term learns within-region relationships while the unary term learns salient boundaries. However, the two terms are tightly coupled in the non-local block, which hinders the learning of each. Based on these findings, we present the disentangled non-local block, where the two terms are decoupled to facilitate learning for both terms. We demonstrate the effectiveness of the decoupled design on various tasks, such as semantic segmentation on Cityscapes, ADE20K and PASCAL Context, object detection on COCO, and action recognition on Kinetics.
In this paper, we propose a novel data augmentation strategy named Cut-Thumbnail, that aims to improve the shape bias of the network. We reduce an image to a certain size and replace the random region of the original image with the reduced image. The generated image not only retains most of the original image information but also has global information in the reduced image. We call the reduced image as thumbnail. Furthermore, we find that the idea of thumbnail can be perfectly integrated with Mixed Sample Data Augmentation, so we put one image's thumbnail on another image while the ground truth labels are also mixed, making great achievements on various computer vision tasks. Extensive experiments show that Cut-Thumbnail works better than state-of-the-art augmentation strategies across classification, fine-grained image classification, and object detection. On ImageNet classification, ResNet-50 architecture with our method achieves 79.21\% accuracy, which is more than 2.8\% improvement on the baseline.
Automatic pseudo-labeling is a powerful tool to tap into large amounts of sequential unlabeled data. It is specially appealing in safety-critical applications of autonomous driving, where performance requirements are extreme, datasets are large, and manual labeling is very challenging. We propose to leverage sequences of point clouds to boost the pseudolabeling technique in a teacher-student setup via training multiple teachers, each with access to different temporal information. This set of teachers, dubbed Concordance, provides higher quality pseudo-labels for student training than standard methods. The output of multiple teachers is combined via a novel pseudo label confidence-guided criterion. Our experimental evaluation focuses on the 3D point cloud domain and urban driving scenarios. We show the performance of our method applied to 3D semantic segmentation and 3D object detection on three benchmark datasets. Our approach, which uses only 20% manual labels, outperforms some fully supervised methods. A notable performance boost is achieved for classes rarely appearing in training data.
Foundational models, pretrained on a large scale, have demonstrated substantial success across non-medical domains. However, training these models typically requires large, comprehensive datasets, which contrasts with the smaller and more heterogeneous datasets common in biomedical imaging. Here, we propose a multi-task learning strategy that decouples the number of training tasks from memory requirements. We trained a Universal bioMedical PreTrained model (UMedPT) on a multi-task database including tomographic, microscopic, and X-ray images, with various labelling strategies such as classification, segmentation, and object detection. The UMedPT foundational model outperformed ImageNet pretraining and the previous state-of-the-art models. For tasks related to the pretraining database, it maintained its performance with only 1% of the original training data and without fine-tuning. For out-of-domain tasks it required not more than 50% of the original training data. In an external independent validation imaging features extracted using UMedPT proved to be a new standard for cross-center transferability.
Large multimodal models such as Stable Diffusion can generate, detect, and classify new visual concepts after fine-tuning just a single word embedding. Do models learn similar words for the same concepts (i.e. <orange-cat> = orange + cat)? We conduct a large-scale analysis on three state-of-the-art models in text-to-image generation, open-set object detection, and zero-shot classification, and find that new word embeddings are model-specific and non-transferable. Across 4,800 new embeddings trained for 40 diverse visual concepts on four standard datasets, we find perturbations within an $\epsilon$-ball to any prior embedding that generate, detect, and classify an arbitrary concept. When these new embeddings are spliced into new models, fine-tuning that targets the original model is lost. We show popular soft prompt-tuning approaches find these perturbative solutions when applied to visual concept learning tasks, and embeddings for visual concepts are not transferable. Code for reproducing our work is available at: https://visual-words.github.io.
Three-dimensional (3D) reconstruction from two-dimensional images is an active research field in computer vision, with applications ranging from navigation and object tracking to segmentation and three-dimensional modeling. Traditionally, parametric techniques have been employed for this task. However, recent advancements have seen a shift towards learning-based methods. Given the rapid pace of research and the frequent introduction of new image matching methods, it is essential to evaluate them. In this paper, we present a comprehensive evaluation of various image matching methods using a structure-from-motion pipeline. We assess the performance of these methods on both in-domain and out-of-domain datasets, identifying key limitations in both the methods and benchmarks. We also investigate the impact of edge detection as a pre-processing step. Our analysis reveals that image matching for 3D reconstruction remains an open challenge, necessitating careful selection and tuning of models for specific scenarios, while also highlighting mismatches in how metrics currently represent method performance.
Conventional cameras capture image irradiance on a sensor and convert it to RGB images using an image signal processor (ISP). The images can then be used for photography or visual computing tasks in a variety of applications, such as public safety surveillance and autonomous driving. One can argue that since RAW images contain all the captured information, the conversion of RAW to RGB using an ISP is not necessary for visual computing. In this paper, we propose a novel $\rho$-Vision framework to perform high-level semantic understanding and low-level compression using RAW images without the ISP subsystem used for decades. Considering the scarcity of available RAW image datasets, we first develop an unpaired CycleR2R network based on unsupervised CycleGAN to train modular unrolled ISP and inverse ISP (invISP) models using unpaired RAW and RGB images. We can then flexibly generate simulated RAW images (simRAW) using any existing RGB image dataset and finetune different models originally trained for the RGB domain to process real-world camera RAW images. We demonstrate object detection and image compression capabilities in RAW-domain using RAW-domain YOLOv3 and RAW image compressor (RIC) on snapshots from various cameras. Quantitative results reveal that RAW-domain task inference provides better detection accuracy and compression compared to RGB-domain processing. Furthermore, the proposed \r{ho}-Vision generalizes across various camera sensors and different task-specific models. Additional advantages of the proposed $\rho$-Vision that eliminates the ISP are the potential reductions in computations and processing times.
In this work, we present a novel data-driven method for robust 6DoF object pose estimation from a single RGBD image. Unlike previous methods that directly regressing pose parameters, we tackle this challenging task with a keypoint-based approach. Specifically, we propose a deep Hough voting network to detect 3D keypoints of objects and then estimate the 6D pose parameters within a least-squares fitting manner. Our method is a natural extension of 2D-keypoint approaches that successfully work on RGB based 6DoF estimation. It allows us to fully utilize the geometric constraint of rigid objects with the extra depth information and is easy for a network to learn and optimize. Extensive experiments were conducted to demonstrate the effectiveness of 3D-keypoint detection in the 6D pose estimation task. Experimental results also show our method outperforms the state-of-the-art methods by large margins on several benchmarks. Code and video are available at https://github.com/ethnhe/PVN3D.git.
Removing rain streaks from rainy images is necessary for many tasks in computer vision, such as object detection and recognition. It needs to address two mutually exclusive objectives: removing rain streaks and reserving realistic details. Balancing them is critical for de-raining methods. We propose an end-to-end network, called dual-task de-raining network (DTDN), consisting of two sub-networks: generative adversarial network (GAN) and convolutional neural network (CNN), to remove rain streaks via coordinating the two mutually exclusive objectives self-adaptively. DTDN-GAN is mainly used to remove structural rain streaks, and DTDN-CNN is designed to recover details in original images. We also design a training algorithm to train these two sub-networks of DTDN alternatively, which share same weights but use different training sets. We further enrich two existing datasets to approximate the distribution of real rain streaks. Experimental results show that our method outperforms several recent state-of-the-art methods, based on both benchmark testing datasets and real rainy images.
Contrastive learning methods have significantly narrowed the gap between supervised and unsupervised learning on computer vision tasks. In this paper, we explore their application to geo-located datasets, e.g. remote sensing, where unlabeled data is often abundant but labeled data is scarce. We first show that due to their different characteristics, a non-trivial gap persists between contrastive and supervised learning on standard benchmarks. To close the gap, we propose novel training methods that exploit the spatio-temporal structure of remote sensing data. We leverage spatially aligned images over time to construct temporal positive pairs in contrastive learning and geo-location to design pre-text tasks. Our experiments show that our proposed method closes the gap between contrastive and supervised learning on image classification, object detection and semantic segmentation for remote sensing. Moreover, we demonstrate that the proposed method can also be applied to geo-tagged ImageNet images, improving downstream performance on various tasks. Project Webpage can be found at this link geography-aware-ssl.github.io.
The semantic segmentation (SS) task aims to create a dense classification by labeling at the pixel level each object present on images. Convolutional neural network (CNN) approaches have been widely used, and exhibited the best results in this task. However, the loss of spatial precision on the results is a main drawback that has not been solved. In this work, we propose to use a multi-task approach by complementing the semantic segmentation task with edge detection, semantic contour, and distance transform tasks. We propose that by sharing a common latent space, the complementary tasks can produce more robust representations that can enhance the semantic labels. We explore the influence of contour-based tasks on latent space, as well as their impact on the final results of SS. We demonstrate the effectiveness of learning in a multi-task setting for hourglass models in the Cityscapes, CamVid, and Freiburg Forest datasets by improving the state-of-the-art without any refinement post-processing.
With the burgeoning amount of data of image-text pairs and diversity of Vision-and-Language (V\&L) tasks, scholars have introduced an abundance of deep learning models in this research domain. Furthermore, in recent years, transfer learning has also shown tremendous success in Computer Vision for tasks such as Image Classification, Object Detection, etc., and in Natural Language Processing for Question Answering, Machine Translation, etc. Inheriting the spirit of Transfer Learning, research works in V\&L have devised multiple pretraining techniques on large-scale datasets in order to enhance the performance of downstream tasks. The aim of this article is to provide a comprehensive revision of contemporary V\&L pretraining models. In particular, we categorize and delineate pretraining approaches, along with the summary of state-of-the-art vision-and-language pretrained models. Moreover, a list of training datasets and downstream tasks is supplied to further polish the perspective into V\&L pretraining. Lastly, we decided to take a further step to discuss numerous directions for future research.