To address the problem of insufficient accuracy of the traditional Yolov5 convolutional neural network model under low resolution and high background noise of infrared thermal images, this paper proposes an improved s...
详细信息
Skin cancer is one of the most commonly diagnosed types of cancer and poses a great threat to people's health. By using computer-aided diagnosis, the corresponding lesions of skin cancer can be classified and segm...
详细信息
The restoration of images affected by severe weather conditions such as heavy fog is a trending topic in the field of computervision. Despite the fact that many image dehazing methods have achieved impressive perform...
详细信息
Non-Local Self-Similarity (NLSS) is a widely exploited prior in image denoising algorithms. The first deep Convolutional Neural Networks (CNNs) for image denoising ignored NLSS and were made of a sequence of convoluti...
详细信息
ISBN:
(数字)9783031431487
ISBN:
(纸本)9783031431470;9783031431487
Non-Local Self-Similarity (NLSS) is a widely exploited prior in image denoising algorithms. The first deep Convolutional Neural Networks (CNNs) for image denoising ignored NLSS and were made of a sequence of convolutional layers trained to suppress noise. The first denoising CNNs leveraging NLSS prior were performing non-learnable operations outside the network. Then, pre-defined similarity measures were introduced and finally learnable, but scalar, similarity scores were adopted inside the network. We propose the Self-Similarity Block (SSB), a novel differentiable building block for CNN denoisers to promote the NLSS prior. The SSB is trained in an end-to-end manner within convolutional layers and learns a multivariate similarity score to improve image denoising by combining similar vectors in an activation map. We test SSB on additive white Gaussian noise suppression, and we show it is particularly beneficial when the noise level is high. Remarkably, SSB is mostly effective in image regions presenting repeated patterns, which most benefit from the NLSS prior.
SPECT imaging is one of the main functional imaging methods in the field of medical imaging, playing an important role in the detection and diagnosis of bone metastasis caused by diverse primary tumors. Bone metastasi...
详细信息
With the rapid development of digital technology and deep learning, recovering 3D scene information and reconstructing human bodies from a single image has become a focal point of research in computervision and compu...
详细信息
Using a deep learning process called semantic segmentation, each pixel in an image is given a label or classification. The groups of pixels that make up the distinct categories are identified using it. In several fiel...
详细信息
Deep learning is a sophisticated and adaptable technique that has found widespread use in fields such as natural language processing, machine learning, and computervision. It is one of the most recent deep learning-p...
详细信息
We present Learning to Explain (LTX), a model-agnostic framework designed for providing post -hoc explanations for vision models. The LTX framework introduces an "explainer" model that generates explanation ...
详细信息
ISBN:
(纸本)9798350307887
We present Learning to Explain (LTX), a model-agnostic framework designed for providing post -hoc explanations for vision models. The LTX framework introduces an "explainer" model that generates explanation maps, highlighting the crucial regions that justify the predictions made by the model being explained. To train the explainer, we employ a two -stage process consisting of initial pretraining followed by per-instance finetuning. During both stages of training, we utilize a unique configuration where we compare the explained model's prediction for a masked input with its original prediction for the unmasked input. This approach enables the use of a novel counterfactual objective, which aims to anticipate the model's output using masked versions of the input image. Importantly, the LTX framework is not restricted to a specific model architecture and can provide explanations for both Transformer-based and convolutional models. Through our evaluations, we demonstrate that LTX significantly outperforms the current state-of-the-art in explainability across various metrics. Our code is available at: https://***/LTX-CodelLTX
image restoration is a fundamental task in low-level computervision. Most existing algorithms assume that the input image has a single known degradation type. In reality, images usually contain multiple degradations,...
详细信息
ISBN:
(纸本)9781665468916
image restoration is a fundamental task in low-level computervision. Most existing algorithms assume that the input image has a single known degradation type. In reality, images usually contain multiple degradations, making the restoration challenging. Though recent works restore the multiple degraded images, they assume that the degradation history is known. Obviously, such an ideal assumption often does not hold in real applications. This work proposes a novel restoration framework for multiple degraded images via degradation history estimation. Specifically, we first develop a sequential model to estimate the degradation history, including both the degradation operation chain and the corresponding parameters. By resorting to designed self-attention and cross-attention mechanisms, our method can effectively model the correlation of the input image, degradation operation chain, and parameters. Then, we apply our estimation framework for the multiple degraded image restoration, without requiring the degradation history. Experiment results demonstrate much better performance than existing approaches.
暂无评论