A low-light image enhancement is a highly demanded image processing technique, especially for consumer digital cameras and cameras on mobile phones. In this paper, a gradient-based low-light image enhancement algorithm is proposed. The key is to enhance the gradients of dark region, because the gradients are more sensitive for human visual system than absolute values. In addition, we involve the intensity-range constraints for the image integration. By using the intensity-range constraints, we can integrate the output image with enhanced gradients preserving the given gradient information while enforcing the intensity range of the output image within a certain intensity range. Experiments demonstrate that the proposed gradientbased low-light image enhancement can effectively enhance the low-light images.
Lifelong learning aims to train a highly expressive model for a new task while retaining all knowledge for previous tasks. However, many practical scenarios do not always require the system to remember all of the past knowledge. Instead, ethical considerations call for selective and proactive forgetting of undesirable knowledge in order to prevent privacy issues and data leakage. In this paper, we propose a new framework for lifelong learning, called Learning with Selective Forgetting, which is to update a model for the new task with forgetting only the selected classes of the previous tasks while maintaining the rest. The key is to introduce a class-specific synthetic signal called mnemonic code. The codes are "watermarked" on all the training samples of the corresponding classes when the model is updated for a new task. This enables us to forget arbitrary classes later by only using the mnemonic codes without using the original data. Experiments on common benchmark datasets demonstrate the remarkable superiority of the proposed method over several existing methods.
Abstract. A meteorological balloon-borne cloud sensor called the Cloud Particle Sensor (CPS) has been developed. The CPS is equipped with a diode laser at ~ 790 nm and two photo detectors, with a polarization plate in front of one of the detectors, to count the number of particles per second and to obtain the cloud phase information (i.e. liquid, ice, or mixed). The lower detection limit for particle size was evaluated in laboratory experiments as ~ 2 μm diameter for water droplets. For the current model the output voltage often saturates for water droplets with diameter equal to or greater than ~ 80 μm. The upper limit of the directly measured particle number concentration is ~ 2 cm−3 (2 × 103L−1), which is determined by the volume of the detection area of the instrument. In a cloud layer with a number concentration higher than this value, particle signal overlap and multiple scattering of light occur within the detection area, resulting in a counting loss, though a partial correction may be possible using the particle signal width data. The CPS is currently interfaced with either a Meisei RS-06G radiosonde or a Meisei RS-11G radiosonde that measures vertical profiles of temperature, relative humidity, height, pressure, and horizontal winds. Twenty-five test flights have been made between 2012 and 2015 at midlatitude and tropical sites. In this paper, results from four flights are discussed in detail. A simultaneous flight of two CPSs with different instrumental configurations confirmed the robustness of the technique. At a midlatitude site, a profile containing, from low to high altitude, water clouds, mixed phase clouds, and ice clouds was successfully obtained. In the tropics, vertically thick cloud layers in the middle to upper troposphere and vertically thin cirrus layers in the upper troposphere were successfully detected in two separate flights. The data quality is much better at night, dusk and dawn than during the daytime because strong sunlight affects the measurements of scattered light.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.