Semantic segmentation is a computer vision task that associates a label with each pixel in an image. Modern approaches tend to introduce class embeddings into semantic segmentation for deeply utilizing category semantics, and regard supervised class masks as final predictions. In this paper, we explore the mechanism of class embeddings and have an insight that more explicit and meaningful class embeddings can be generated based on class masks purposely. Following this observation, we propose ECENet, a new segmentation paradigm, in which class embeddings are obtained and enhanced explicitly during interacting with multi-stage image features. Based on this, we revisit the traditional decoding process and explore inverted information flow between segmentation masks and class embeddings. Furthermore, to ensure the discriminability and informativity of features from backbone, we propose a Feature Reconstruction module, which combines intrinsic and diverse branches together to ensure the concurrence of diversity and redundancy in features. Experiments show that our ECENet outperforms its counterparts on the ADE20K dataset with much less computational cost and achieves new state-of-the-art results on PASCAL-Context dataset. The code will be released at https://gitee.com/mindspore/models and https://github.com/Carol-lyh/ECENet.
Contrastive learning is a powerful way of learning multimodal representations across various domains such as image-caption retrieval and audio-visual representation learning. In this work, we investigate if these findings generalize to the domain of music videos. Specifically, we create a dual en-coder for the audio and video modalities and train it using a bidirectional contrastive loss. For the experiments, we use an industry dataset containing 550 000 music videos as well as the public Million Song Dataset, and evaluate the quality of learned representations on the downstream tasks of music tagging and genre classification. Our results indicate that pre-trained networks without contrastive fine-tuning outperform our contrastive learning approach when evaluated on both tasks. To gain a better understanding of the reasons contrastive learning was not successful for music videos, we perform a qualitative analysis of the learned representations, revealing why contrastive learning might have difficulties uniting embeddings from two modalities. Based on these findings, we outline possible directions for future work. To facilitate the reproducibility of our results, we share our code and the pre-trained model.
A diversified dataset is crucial for training a well-generalized supervised computer vision algorithm. However, in the field of microbiology, generation and annotation of a diverse dataset including field-taken images are time consuming, costly, and in some cases impossible. Image to image translation frameworks allow us to diversify the dataset by transferring images from one domain to another. However, most existing image translation techniques require a paired dataset (original image and its corresponding image in the target domain), which poses a significant challenge in collecting such datasets. In addition, the application of these image translation frameworks in microbiology is rarely discussed. In this study, we aim to develop an unpaired GAN-based (Generative Adversarial Network) image to image translation model for microbiological images, and study how it can improve generalization ability of object detection models. In this paper, we present an unpaired and unsupervised image translation model to translate laboratory-taken microbiological images to field images, building upon the recent advances in GAN networks and Perceptual loss function. We propose a novel design for a GAN model, BioGAN, by utilizing Adversarial and Perceptual loss in order to transform high level features of laboratory-taken images into field images, while keeping their spatial features. The contribution of Adversarial and Perceptual loss in the generation of realistic field images were studied. We used the synthetic field images, generated by BioGAN, to train an object-detection framework, and compared the results with those of an object-detection framework trained with laboratory images; this resulted in up to 68.1% and 75.3% improvement on F1-score and mAP, respectively. Codes is publicly available at https://github.com/Kahroba2000/BioGAN.
Contact is at the core of robotic manipulation. At times, it is desired (e.g. manipulation and grasping), and at times, it is harmful (e.g. when avoiding obstacles). However, traditional path planning algorithms focus solely on collision-free paths, limiting their applicability in contact-rich tasks. To address this limitation, we propose the domain of Language-Conditioned Path Planning, where contact-awareness is incorporated into the path planning problem. As a first step in this domain, we propose Language-Conditioned Collision Functions (LACO) a novel approach that learns a collision function using only a single-view image, language prompt, and robot configuration. LACO predicts collisions between the robot and the environment, enabling flexible, conditional path planning without the need for manual object annotations, point cloud data, or ground-truth object meshes. In both simulation and the real world, we demonstrate that LACO can facilitate complex, nuanced path plans that allow for interaction with objects that are safe to collide, rather than prohibiting any collision.
A central challenge of video prediction lies where the system has to reason the objects' future motions from image frames while simultaneously maintaining the consistency of their appearances across frames. This work introduces an end-to-end trainable two-stream video prediction framework, Motion-Matrix-based Video Prediction (MMVP), to tackle this challenge. Unlike previous methods that usually handle motion prediction and appearance maintenance within the same set of modules, MMVP decouples motion and appearance information by constructing appearance-agnostic motion matrices. The motion matrices represent the temporal similarity of each and every pair of feature patches in the input frames, and are the sole input of the motion prediction module in MMVP. This design improves video prediction in both accuracy and efficiency, and reduces the model size. Results of extensive experiments demonstrate that MMVP outperforms state-of-the-art systems on public data sets by non-negligible large margins (about 1 db in PSNR, UCF Sports) in significantly smaller model sizes (84% the size or smaller).
Current 3D open-vocabulary scene understanding methods mostly utilize well-aligned 2D images as the bridge to learn 3D features with language. However, applying these approaches becomes challenging in scenarios where 2D images are absent. In this work, we introduce a completely new pipeline, namely, OpenIns3D, which requires no 2D image inputs, for 3D open-vocabulary scene understanding at the instance level. The OpenIns3D framework employs a "Mask-Snap-Lookup" scheme. The "Mask" module learns class-agnostic mask proposals in 3D point clouds. The "Snap" module generates synthetic scene-level images at multiple scales and leverages 2D vision language models to extract interesting objects. The "Lookup" module searches through the outcomes of "Snap" with the help of Mask2Pixel maps, which contain the precise correspondence between 3D masks and synthetic images, to assign category names to the proposed masks. This 2D input-free, easy-to-train, and flexible approach achieved state-of-the-art results on a wide range of indoor and outdoor datasets with a large margin. Furthermore, OpenIns3D allows for effortless switching of 2D detectors without re-training. When integrated with state-of-the-art 2D open-world models such as ODISE and GroundingDINO, superb results are observed on open-vocabulary instance segmentation. When integrated with LLM-powered 2D models like LISA, it demonstrates a remarkable capacity to process highly complex text queries, including those that require intricate reasoning and world knowledge. Project page: https://zheninghuang.github.io/OpenIns3D/
Deep learning techniques for medical image analysis usually suffer from the domain shift between source and target data. Most existing works focus on unsupervised domain adaptation (UDA). However, in practical applications, privacy issues are much more severe. For example, the data of different hospitals have domain shifts due to equipment problems, and data of the two domains cannot be available simultaneously because of privacy. In this challenge defined as Source-Free UDA, the previous UDA medical methods are limited. Although a variety of medical source-free unsupervised domain adaption (MSFUDA) methods have been proposed, we found they fall into an over-fitting dilemma called "longer training, worse performance." Therefore, we propose the Stable Learning (SL) strategy to address the dilemma. SL is a scalable method and can be integrated with other research, which consists of Weight Consolidation and Entropy Increase. First, we apply Weight Consolidation to retain domain-invariant knowledge and then we design Entropy Increase to avoid over-learning. Comparative experiments prove the effectiveness of SL. We also have done extensive ablation experiments. Besides, We will release codes including a variety of MSFUDA methods.
Stroke lesion volume is a key radiologic measurement for assessing the prognosis of Acute Ischemic Stroke (AIS) patients, which is challenging to be automatically measured on Non-Contrast CT (NCCT) scans. Recent diffusion probabilistic models have shown potentials of being used for image segmentation. In this paper, a novel Synchronous image-label Diffusion Probability Model (SDPM) is proposed for stroke lesion segmentation on NCCT using Markov diffusion process. The proposed SDPM is fully based on a Latent Variable Model (LVM), offering a complete probabilistic elaboration. An additional net-stream, parallel with a noise prediction stream, is introduced to obtain initial noisy label estimates for efficiently inferring the final labels. By optimizing the specified variational boundaries, the trained model can infer multiple label estimates for reference given the input images with noises. The proposed model was assessed on three stroke lesion datasets including one public and two private datasets. Compared to several U-net and transformer-based segmentation methods, our proposed SDPM model is able to achieve state-of-the-art performance. The code is publicly available.
Lesion volume is an important predictor for prognosis in breast cancer. We make a step towards a more accurate lesion volume measurement on digital mammograms by developing a model that allows to estimate lesion volumes on processed mammograms, which are the images routinely used by radiologists in clinical practice as well as in breast cancer screening and are available in medical centers. Processed mammograms are obtained from raw mammograms, which are the X-ray data coming directly from the scanner, by applying certain vendor-specific non-linear transformations. At the core of our volume estimation method is a physics-based algorithm for measuring lesion volumes on raw mammograms. We subsequently extend this algorithm to processed mammograms via a deep learning image-to-image translation model that produces synthetic raw mammograms from processed mammograms in a multi-vendor setting. We assess the reliability and validity of our method using a dataset of 1778 mammograms with an annotated mass. Firstly, we investigate the correlations between lesion volumes computed from mediolateral oblique and craniocaudal views, with a resulting Pearson correlation of 0.93 [95% confidence interval (CI) 0.92 - 0.93]. Secondly, we compare the resulting lesion volumes from true and synthetic raw data, with a resulting Pearson correlation of 0.998 [95% CI 0.998 - 0.998] . Finally, for a subset of 100 mammograms with a malign mass and concurrent MRI examination available, we analyze the agreement between lesion volume on mammography and MRI, resulting in an intraclass correlation coefficient of 0.81 [95% CI 0.73 - 0.87] for consistency and 0.78 [95% CI 0.66 - 0.86] for absolute agreement. In conclusion, we developed an algorithm to measure mammographic lesion volume that reached excellent reliability and good validity, when using MRI as ground truth.
In this work, the novel Image Transformation Sequence Retrieval (ITSR) task is presented, in which a model must retrieve the sequence of transformations between two given images that act as source and target, respectively. Given certain characteristics of the challenge such as the multiplicity of a correct sequence or the correlation between consecutive steps of the process, we propose a solution to ITSR using a general model-based Reinforcement Learning such as Monte Carlo Tree Search (MCTS), which is combined with a deep neural network. Our experiments provide a benchmark in both synthetic and real domains, where the proposed approach is compared with supervised training. The results report that a model trained with MCTS is able to outperform its supervised counterpart in both the simplest and the most complex cases. Our work draws interesting conclusions about the nature of ITSR and its associated challenges.