RESUMEN
Analog in-memory computing synaptic devices are widely studied for efficient implementation of deep learning. However, synaptic devices based on resistive memory have difficulties implementing on-chip training due to the lack of means to control the amount of resistance change and large device variations. To overcome these shortcomings, silicon complementary metal-oxide semiconductor (Si-CMOS) and capacitor-based charge storage synapses are proposed, but it is difficult to obtain sufficient retention time due to Si-CMOS leakage currents, resulting in a deterioration of training accuracy. Here, a novel 6T1C synaptic device using only n-type indium gaIlium zinc oxide thin film transistor (IGZO TFT) with low leakage current and a capacitor is proposed, allowing not only linear and symmetric weight update but also sufficient retention time and parallel on-chip training operations. In addition, an efficient and realistic training algorithm to compensate for any remaining device non-idealities such as drifting references and long-term retention loss is proposed, demonstrating the importance of device-algorithm co-optimization.
RESUMEN
One of the current issues for picture archiving and communication systems (PACS) is extending retrieval technologies to deal with multimedia information. This is particularly important for medical applications that assist in diagnostic processes and pathology studies. Accordingly, this paper presents a new approach to content-based image retrieval (CBIR) for a clinical ultrasound image database (DB). The proposed algorithm consists of two stages so as to maximize the retrieval efficiency. In the first stage, a coarse retrieval is performed using the statistical characteristics of the wavelet coefficients that narrow the search by eliminating up to 70% of the total DB images. In the second stage, a fine retrieval is carried out using the Legendre moment of the global histogram pdf on the reduced image set preretrieved by the coarse retrieval. When tested on an abdominal ultrasound image DB and compared with various other methods, the proposed algorithm gave promising results for applying CBIR to clinical ultrasound images.
