Low-light images suffer from low contrast and unclear details, which not only reduces the available information for humans but limits the application of computer vision algorithms. Among the existing enhancement techniques, Retinex-based and learning-based methods are under the spotlight today. In this paper, we bridge the gap between the two methods. First, we propose a novel “generative” strategy for Retinex decomposition, by which the decomposition is cast as a generative problem. Second, based on the strategy, a unified deep framework is proposed to estimate the latent components and perform low-light image enhancement. Third, our method can weaken the coupling relationship between the two components while performing Retinex decomposition. Finally, the RetinexDIP performs Retinex decomposition without any external images, and the estimated illumination can be easily adjusted and is used to perform enhancement. The proposed method is compared with ten state-of-the-art algorithms on seven public datasets, and the experimental results demonstrate the superiority of our method. Code is available at: https://github.com/zhaozunjin/RetinexDIP .

Realization of brain-like computer has always been human’s ultimate dream. Today, the possibility of having this dream come true has been significantly boosted due to the advent of several emerging non-volatile memory devices. Within these innovative technologies, phase-change memory device has been commonly regarded as the most promising candidate to imitate the biological brain, owing to its excellent scalability, fast switching speed, and low energy consumption. In this context, a detailed review concerning the physical principles of the neuromorphic circuit using phase-change materials as well as a comprehensive introduction of the currently available phase-change neuromorphic prototypes becomes imperative for scientists to continuously progress the technology of artificial neural networks. In this paper, we first present the biological mechanism of human brain, followed by a brief discussion about physical properties of phase-change materials that recently receive a widespread application on non-volatile memory field. We then survey recent research on different types of neuromorphic circuits using phase-change materials in terms of their respective geometrical architecture and physical schemes to reproduce the biological events of human brain, in particular for spike-time-dependent plasticity. The relevant virtues and limitations of these devices are also evaluated. Finally, the future prospect of the neuromorphic circuit based on phase-change technologies is envisioned.

Non-volatile storage memory is widely considered to be one of the most promising candidates to replace dynamic random access memory and even static random access memory. It has recently received particular attention because of its great potential for brain-like neuromorphic applications. Phase-change materials, also known as Chalcogenide alloys, exhibit several especially advantageous traits for non-volatile applications. These include scalability, fast switching speeds, low switching energy and outstanding thermal stability. As a result, most research to date has sought to identify electrical applications for phase-change materials in relation to phase-change random access memory, phase-change memristors and phase-change neuro networks, while overlooking their potential for non-volatile photonic applications. To address this issue, we provide a comprehensive review that examines the remarkable physical properties of phase-change materials for photonic applications, together with emerging phase-change photonic devices. The review begins by presenting the atomic structure and physical properties of phase-change materials, followed by an elaboration of the issues that phase-change materials are currently facing and the strategies being developed to overcome them. The current state-of-the-art and technical challenges confronting phase-change materials in relation to non-volatile photonic applications, such as phase-change photonic memory, phase-change photonic neuro-networks, phase-change metasurfaces and phase-change color displays are then considered. The review concludes by discussing the outlook for successfully implementing phase-change materials in emerging photonic domains.

Memristor is a fundamental circuit element in addition to resistor, capacitor, and inductor. As it can remember its resistance state even encountering a power off, memristor has recently received widespread applications from non-volatile memory to neural networks. The current memristor family mainly comprises resistive memristor, polymeric memristor, ferroelectric memristor, manganite memristor, resonant-tunneling diode memristor, and spintronic memristor in terms of the materials the device is made of. In order to help researcher better understand the physical principles of the memristor, and thus to provide a promising prospect for memristor devices, this paper presents an overview of memristor materials properties, switching mechanisms, and potential applications. The performance comparison among different memristor members is also given.

A systematic design of practicable media suitable for rewritable, ultrahigh density (>;1Tbit/sq.in.), high data rate (>;1Mbit/s/tip) scanning-probe phase-change memories is presented. The basic design requirements were met by a Si/TiN/Ge₂Sb₂Te₅ (GST)/diamond-like carbon structure, with properly tailored electrical and thermal conductivities. Various alternatives for providing rewritability were investigated. In the first case, amorphous marks were written into a crystalline starting phase and subsequently erased by recrystallization, as in other already established phase-change memory technologies. Results imply that this approach is also appropriate for probe-based memories. However, experimentally, the successful writing of amorphous bits using scanning electrical probes has not been widely reported. In light of this, a second approach has been studied, that of writing crystalline bits in an amorphous starting matrix, with subsequent erasure by reamorphization. With conventional phase-change materials, such as continuous films of GST, this approach invariably leads to the formation of a crystalline “halo” surrounding the erased (reamorphized) region, with severe adverse consequences on the achievable density. Suppression of the “halo” was achieved using patterned media or slow-growth phase-change media, with the latter seemingly more viable.