An official website of the United States government
Most neural networks assume that input images have a fixed number of channels (three for RGB images). However, there are many settings where the number of channels may vary, such as microscopy images where the number of channels changes depending on instruments and experimental goals. Yet, there has not been a systemic attempt to create and evaluate neural networks that are invariant to the number and type of channels. As a result, trained models remain specific to individual studies and are hardly reusable for other microscopy settings. In this paper, we present a benchmark for investigating channel-adaptive models in microscopy imaging, which consists of 1) a dataset of varied-channel single-cell images, and 2) a biologically relevant evaluation framework. In addition, we adapted several existing techniques to create channel-adaptive models and compared their performance on this benchmark to fixed-channel, baseline models. We find that channel-adaptive models can generalize better to out-of-domain tasks and can be computationally efficient. We contribute a curated dataset and an evaluation API to facilitate objective comparisons in future research and applications.
more »
« less
This content will become publicly available on December 12, 2025
Enhancing feature diversity boosts channel-adaptive vision transformers
Multi-Channel Imaging (MCI) contains an array of challenges for encoding useful feature representations not present in traditional images. For example, images from two different satellites may both contain RGB channels, but the remaining channels can be different for each imaging source. Thus, MCI models must support a variety of channel configurations at test time. Recent work has extended traditional visual encoders for MCI, such as Vision Transformers (ViT), by supplementing pixel information with an encoding representing the channel configuration. However, these methods treat each channel equally, i.e., they do not consider the unique properties of each channel type, which can result in needless and potentially harmful redundancies in the learned features. For example, if RGB channels are always present, the other channels can focus on extracting information that cannot be captured by the RGB channels. To this end, we propose DiChaViT, which aims to enhance the diversity in the learned features of MCI-ViT models. This is achieved through a novel channel sampling strategy that encourages the selection of more distinct channel sets for training. Additionally, we employ regularization and initialization techniques to increase the likelihood that new information is learned from each channel. Many of our improvements are architecture agnostic and can be incorporated into new architectures as they are developed. Experiments on both satellite and cell microscopy datasets, CHAMMI, JUMP-CP, and So2Sat, report DiChaViT yields a 1.5 - 5.0% gain over the state-of-the-art. Our code is publicly available at https://github.com/chaudatascience/diversechannelvit.
more »
« less
- Award ID(s):
- 2134696
- PAR ID:
- 10631605
- Publisher / Repository:
- Advances in Neural Information Processing Systems
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Recent advances in self-supervised pre-training of foundation models for natural images have made them a popular choice for various visual systems and applications. Self-supervised strategies are also promising in non-RGB scientific imaging domains such as in biology, medical and satellite imagery, but their broader application is hampered by heterogeneity in channel composition and semantics between relevant datasets: two datasets may contain different numbers of channels, and these may reveal distinct aspects of an object or scene. Recent works on channel adaptive strategies report substantial advantages for those that account for variable channel compositions without sacrificing the ability to jointly encode channels; yet, how these strategies behave at scale remains unclear. We here show that, surprisingly, trained across large-scale datasets, independent-encoding of channels outperforms jointencoding methods by a substantial margin. We validate this result along an extensive set of experiments on various datasets from cell microscopy to geospatial imagery. Our DINO BoC approach sets a new state-of-the-art across challenging benchmarks, including generalization to out-of-distribution tasks and unseen channel combinations. We open-source code and model weights for a new general-purpose feature extractor for fluorescent microscopymore » « less
-
Major semantic segmentation approaches are designed for RGB color images, which is interpolated from raw Bayer images. The use of RGB images on the one hand provides abundant scene color information. On the other hand, RGB images are easily observable for human users to understand the scene. The RGB color continuity also facilitates researchers to design segmentation algorithms, which becomes unnecessary in end-to-end learning. More importantly, the use of 3 channels adds extra storage and computation burden for neural networks. In contrast, the raw Bayer images can reserve the primitive color information in the largest extent with just a single channel. The compact design of Bayer pattern not only potentially increases a higher segmentation accuracy because of avoiding interpolation, but also significantly decreases the storage requirement and computation time in comparison with standard R, G, B images. In this paper, we propose BayerSeg-Net to segment single channel raw Bayer image directly. Different from RGB color images that already contain neighboring context information during ISP color interpolation, each pixel in raw Bayer images does not contain any context clues. Based on Bayer pattern properties, BayerSeg-Net assigns dynamic attention on Bayer images' spectral frequency and spatial locations to mitigate classification confusion, and proposes a re-sampling strategy to capture both global and local contextual information. We demonstrate the usability of raw Bayer images in segmentation tasks and the efficiency of BayerSeg-Net on multiple datasets.more » « less
-
Semantic segmentation methods are typically designed for RGB color images, which are interpolated from raw Bayer images. While RGB images provide abundant color information and are easily understood by humans, they also add extra storage and computational burden for neural networks. On the other hand, raw Bayer images preserve primitive color information with a single channel, potentially increasing segmentation accuracy while significantly decreasing storage and computation time. In this paper, we propose RawSeg-Net to segment single-channel raw Bayer images directly. Different from RGB images that already contain neighboring context information during ISP color interpolation, each pixel in raw Bayer images does not contain any context clues. Based on Bayer pattern properties, RawSeg-Net assigns dynamic attention on Bayer images' spectral frequency and spatial locations to mitigate classification confusion, and proposes a re-sampling strategy to capture both global and local contextual information.more » « less
-
Global warming is one of the world’s most pressing issues. The study of its effects on the polar ice caps and other arctic environments, however, can be hindered by the often dangerous and difficult to navigate terrain found there. Multi-terrain autonomous vehicles can assist researchers by providing a mobile platform on which to collect data in these harsh environments while avoiding any risk to human life and speeding up the research process. The mechanical design and ultimate efficacy of these autonomous robotic vehicles depends largely on the specific missions they are deployed for, but terrain conditions can vary wildly geographically as well as seasonally, making mission planning for these unmanned vehicles more difficult. This paper proposes the use of various UNet-based neural network architectures to generate digital elevation maps from satellite images, and explores and compares their efficacy on a single set of training and validation datasets generated from satellite imagery. These digital elevation maps generated by the model could be used by researchers not only to track the change in arctic topography over time, but to quickly provide autonomous exploratory research rovers with the topographical information necessary to decide on optimal paths during the mission. This paper analyzes different model architectures and training schemes: a traditional UNet, a traditional UNet with data augmentation, a UNet with a single active skip-layer vision transformer (ViT), and a UNet with multiple active skip-layer ViT. Each model was trained on a dataset of satellite images and corresponding digital elevation maps of Ellesmere Island, Canada. Utilizing ViTs did not demonstrate a significant improvement in UNet performance, though this could change with longer training. This paper proposes opportunities to improve performance for these neural networks, as well as next steps for further research, including improving the diversity of images in the dataset, generating a testing dataset from a completely different geographic location, and allowing the models more time to train.more » « less
