This paper proposes facial recognition approach on variations in outdoor lighting conditions using short wavelength infrared (SWIR) images. Conventional visible light (VIS) systems struggle under challenging lighting, such as backlit condition, due to their sensitivity to ambient light. Recently, SWIR imaging has emerged as a promising alternative, offering robustness against such conditions. Despite its advantages, SWIR imaging faces challenges, including higher noise levels, limited high-intensity illumination, and differences in image characteristics compared to VIS images. These challenges hinder the integration of SWIR-based systems with existing VIS databases. To address these issues, this study proposes a SWIR-to-VIS image translation method to improve VIS-SWIR face recognition performance. The proposed method relies solely on recognition loss for training the image translator, eliminating the need for other loss functions. This novel approach ensures compatibility with existing VIS-based databases while taking advantage of SWIR’s robustness to ambient light. Experimental evaluations demonstrate the efficacy of the proposed method in achieving accurate face recognition under severe outdoor lighting conditions.

Low-Light Image Enhancement (LLIE) is crucial for improving both human perception and computer vision tasks. This paper addresses two challenges in zero-reference LLIE: obtaining perceptually ’good’ images using the Contrastive Language-Image Pre-Training (CLIP) model and maintaining computational efficiency for high-resolution images. We propose CLIP-Utilized Reinforcement learning-based Visual image Enhancement (CURVE). CURVE employs a simple image processing module which adjusts global image tone based on Bézier curve and estimates its processing parameters iteratively. The estimator is trained by reinforcement learning with rewards designed using CLIP text embeddings. Experiments on low-light and multi-exposure datasets demonstrate the performance of CURVE in terms of enhancement quality and processing speed compared to conventional methods.

This paper rethinks image histogram matching (HM) and proposes a differentiable and parametric HM preprocessing for a downstream classifier. Convolutional neural networks have demonstrated remarkable achievements in classification tasks. However, they often exhibit degraded performance on low-contrast images captured under adverse weather conditions. To maintain classifier performance under low-contrast images, histogram equalization (HE) is commonly used. HE is a special case of HM using a uniform distribution as a target pixel value distribution. In this paper, we focus on the shape of the target pixel value distribution. Compared to a uniform distribution, a single, well-designed distribution could have potential to improve the performance of the downstream classifier across various adverse weather conditions. Based on this hypothesis, we propose a differentiable and parametric HM that optimizes the target distribution using the loss function of the downstream classifier. This method addresses pixel value imbalances by transforming input images with arbitrary distributions into a target distribution optimized for the classifier. Our HM is trained on only normal weather images using the classifier. Experimental results show that a classifier trained with our proposed HM outperforms conventional preprocessing methods under adverse weather conditions.

We propose a target driven adaptive (TDA) loss to enhance the performance of infrared small target detection (IRSTD). Prior works have used loss functions, such as binary cross-entropy loss and IoU loss, to train segmentation models for IRSTD. Minimizing these loss functions guides models to extract pixel-level features or global image context. However, they have two issues: improving detection performance for local regions around the targets and enhancing robustness to small scale and low local contrast. To address these issues, the proposed TDA loss introduces a patch-based mechanism, and an adaptive adjustment strategy to scale and local contrast. The proposed TDA loss leads the model to focus on local regions around the targets and pay particular attention to targets with smaller scales and lower local contrast. We evaluate the proposed method on three datasets for IRSTD. The results demonstrate that the proposed TDA loss achieves better detection performance than existing losses on these datasets.

We propose an image-adaptive object detection method for adverse weather conditions such as fog and low-light. Our framework employs differentiable preprocessing filters to perform image enhancement suitable for later-stage object detections. Our framework introduces two differentiable filters: a Bezier curve-based pixel-wise (BPW) filter and a kernel-based local (KBL) filter. These filters unify the functions of classical image processing filters and improve performance of object detection. We also propose a domain-agnostic data augmentation strategy using the BPW filter. Our method does not require data-specific customization of the filter combinations, parameter ranges, and data augmentation. We evaluate our proposed approach, called Enhanced Robustness by Unified Image Processing (ERUP)-YOLO, by applying it to the YOLOv3 detector. Experiments on adverse weather datasets demonstrate that our proposed filters match or exceed the expressiveness of conventional methods and our ERUP-YOLO achieved superior performance in a wide range of adverse weather conditions, including fog and low-light conditions.

This paper proposes a deep feature extractor for iris recognition at arbitrary resolutions. Resolution degradation reduces the recognition performance of deep learning models trained by high-resolution images. Using various-resolution images for training can improve the model’s robustness while sacrificing recognition performance for high-resolution images. To achieve higher recognition performance at various resolutions, we propose a method of resolution-adaptive feature extraction with automatically switching networks. Our framework includes resolution expert modules specialized for different resolution degradations, including down-sampling and out-of-focus blurring. The framework automatically switches them depending on the degradation condition of an input image. Lower-resolution experts are trained by knowledge-distillation from the high-resolution expert in such a manner that both experts can extract common identity features. We applied our framework to three conventional neural network models. The experimental results show that our method enhances the recognition performance at low- resolution in the conventional methods and also maintains their performance at high-resolution.

This paper proposes an efficient iris localization method without using iris segmentation and circle fitting. Conventional iris localization methods first extract iris regions by using semantic segmentation methods such as U-Net. Afterward, the inner and outer iris circles are localized using the traditional circle fitting algorithm. However, this approach requires high-resolution encoder-decoder networks for iris segmentation, so it causes computational costs to be high. In addition, traditional circle fitting tends to be sensitive to noise in input images and fitting parameters, causing the iris recognition performance to be poor. To solve these problems, we propose an iris localization network (ILN), that can directly localize pupil and iris circles with eyelid points from a low-resolution iris image. We also introduce a pupil refinement network (PRN) to improve the accuracy of pupil localization. Experimental results show that the combination of ILN and PRN works in 34.5 ms for one iris image on a CPU, and its localization performance out-performs conventional iris segmentation methods. In addition, generalized evaluation results show that the proposed method has higher robustness for datasets in different domain than other segmentation methods. Furthermore, we also confirm that the proposed ILN and PRN improve the iris recognition accuracy.

This paper proposes a method of inter-orbit change detection for high-resolution synthetic aperture radar (SAR) imagery using a conditional Siamese network. The proposed method introduces a sub-network with a condition, which is satellite orbit information, into the multitask Siamese network (MS-net). To introduce the conditions effectively, the weights sharing in one fully-connected layer after connecting the subnetwork is canceled. These tricks enable the proposed network to learn the absorption of layover effects depending on the orbit, which improves change detection performance. Experiments were conducted for detecting car changes in a parking lot by using Asnaro-2 images captured from five different orbits. Compared with the conventional MS-net, the proposed model improves AUC-ROC by 0.015 on average and is more robust to input orbit combinations.

This paper proposes a framework to train an artifact-free thin cloud removal model using Generative Adversarial Nets (GANs) with thick cloud masks. Satellite images are useful in various applications, however their exploitation is often limited by a presence of clouds. The proposed model can safely remove thin clouds for cloudy images while preserving thick clouds areas without creating undesired artifacts. In order to train the model, we propose a following framework divided in three blocks: generation of thick cloud masks for training images based on texture and spectrum analysis, selection of input-target couples of training images and training of the model using the GAN framework. The use of cloud masks for the training images allows the training of a model for clouds removal, robust to artifact generation in areas with thick clouds. Experimental results show that our model can actually avoid the generation of artifacts, and outperforms the conventional method in terms of SSIM index in testing.

This paper proposes rollable latent space (RLS) for an azimuth invariant synthetic aperture radar (SAR) target recognition. Scarce labeled data and limited viewing direction are critical issues in SAR target recognition. The RLS is a designed space in which rolling of latent features corresponds to 3D rotation of an object. Thus latent features of an arbitrary view can be inferred using those of different views. This characteristic further enables us to augment data from limited viewing in RLS. RLS-based classifiers with and without data augmentation and a conventional classifier trained with target front shots are evaluated over untrained target back shots. Results show that the RLS-based classifier with augmentation improves an accuracy by 30 % compared to the conventional classifier.

This paper presents a method for automatic association between synthetic aperture radar (SAR) and optical images based on zero-shot learning (ZSL). SAR target recognition is an important task in security and defense areas, however, it is still challenging due to insufficient labeled SAR images. In order to solve this problem, we propose a ZSL based SAR target recognition. The conventional ZSL transfers information of observed labels in order to recognize unseen objects. The proposed ZSL replaces the labels with optical images corresponding to SAR targets. The proposed method employs ZSL in conjunction with dimension reduction to match different types of unseen images in a compact feature space with high accuracies. When principle component analysis (PCA) is used for the dimension reduction, the proposed method achieves 0.70 in accuracy while a conventional one shows 0.57.