Current perception models in autonomous driving have become notorious for greatly relying on a mass of annotated data to cover unseen cases and address the long-tail problem. On other hand, learning from unlabeled large-scale collected incrementally self-training powerful recognition received increasing attention may solutions next-generation industry-level robust driving. However, research community generally suffered inadequacy those essential real-world scene data, which hampers future...

Monitoring applications based on wireless camera networks are becoming highly attractive. However, processing and streaming all image data impractical for sensor nodes due to the constraints limited energy, bandwidth, computing resources. In this paper, we propose an energy-efficient compressive transmission scheme, which can provide low-complexity regions of interest extraction, energy-driven quality control, robust communication against packet loss over networks. Performance proposed...

The core challenge of steganography is always how to improve the hidden capacity and concealment. Most current generation-based linguistic methods only consider probability distribution between text characters, emotion topic generated steganographic are uncontrollable. Especially for long texts, generating several sentences related a displaying overall coherence discourse-relatedness can ensure better In this article, we address problem coherent multi-sentence texts concealment, topic-aware...

In this paper we present a parallel-array implementation of new non-restoring square root algorithm (PASQRT). The carry-save adder (CSA) is used in the parallel array. PASQRT has several features unlike other implementations. First, it does not use redundant representation for result. Second, each iteration generates an exact resulting value. Next require any conversion on inputs CSA. And last, precise remainder can be obtained immediately. Furthermore, improved version-a root-select...

Generation-based linguistic steganography is a popular research area of information hiding. The text generative steganographic method based on conditional probability coding the direction that researchers have recently paid attention to. However, in course our experiment, we found secret hiding tends to destroy statistical distribution characteristics original text, which indicates this has problem obvious reduction quality when embedding rate increases, and topic generated texts...

Structural impairments can be detected by monitoring and analyzing dynamic structural responses. In this paper, accurate deformations of multiple vibrating structures are estimated using one off-the-shelf CCD (Charge-coupled Device) camera a novel target configuration computation algorithm. method, camera, which mounted on either an adjacent structure or stationary support, is used to track global movements structure. The method utilizes subpixel feature extraction, Zhang's calibration-based...

By reversing secret images, a novel multi-secret visual cryptography scheme is presented in this paper. During the sharing process, correlative matrices are designed to encode multiple images into two ring shares. Compared with previous ones, proposed makes number of not restricted and has obviously improved pixel expansion relative difference.

Hardware square-root units require large numbers of gates even for iterative implementations. In this paper we present four low-cost high-performance fully-pipelined n-select implementations (nS-Root) based on a non-restoring-remainder square root algorithm. The nS-Root uses parallel array carry-save adders (CSAs). For bit calculation, CSA is used once. This means that the calculations can be fully pipelined. It also n-way root-select technique to speedup calculation. cost/performance...

Most of the existing text generative steganographic methods are based on coding conditional probability distribution each word during generation process, and then selecting specific words according to secret information, so as achieve information hiding. Such have their limitations which may bring potential security risks. Firstly, with increase embedding rate, these models will choose lower probability, reduce quality generated texts; secondly, they can not control semantic expression final...

Computing a distance map (distance transform) is an operation that converts 2D image consisting of black and white pixels to where each pixel has value or pair coordinates represents the location nearest pixel. It basic in processing computer vision fields, used for expanding, shrinking, thinning, segmentation, clustering, computing shape, object reconstruction, etc. This paper examines possibility implementing problem finding efficiently using optical bus. The computational model considered...

Numerous studies have shown that nitric oxide (NO) interacts with human cyclooxygenase (COX); however, conflicting results exist respect to their interactions. Herein, recombinant COX-1 and COX-2 were prepared treated NO donors individually under anaerobic aerobic conditions. The S-nitrosylation detection subsequent kinetic investigations into the arachidonic acid (AA) oxidation of COX enzymes indicate S-nitrosylates both in an oxygen-dependent manner, but enhances only dioxygenase activity...

This paper describes a pipelined processor (named Aizup) design and implementation for the exercise of Computer Architecture/Organization Education at University Aizu. The Aizup, pipeline has four stages deals with data dependency control dependency. Aizup was designed Cadence environment implemented on Xilinx XC4006PC84 FPGA chip. We ask students to processor, perform functional simulations, implement chip, measure chip logic analyzer. course is helpful understand operations processors...

The first quasi-classical trajectory (QCT) calculation for the exothermic reaction Sr + CH 3 Br is carried out based on a constructed London–Eyring–Polanyi–Sato (LEPS) potential energy surface (PES). By QCT calculation, product SrBr vibration distributions are obtained. result in good comparison with experimental one by Keijzer JF et al. [Chem. Phys.207:261, 1996]. Furthermore, angular distribution and rotational alignment also products dominantly forward-scattered effect obvious. Fast...

Addressing the intricate challenge of modeling and re-rendering dynamic scenes, most recent approaches have sought to simplify these complexities using plane-based explicit representations, overcoming slow training time issues associated with methods like Neural Radiance Fields (NeRF) implicit representations. However, straightforward decomposition 4D scenes into multiple 2D representations proves insufficient for high-fidelity complex motions. In response, we present a novel direction-aware...

Modular inversion on large operands is a time-consuming calculation used in elliptic curve cryptography. Its hardware implementation requires extensive resources such as lookup tables and registers. We investigate state-of-the-art modular algorithms evaluate the performance cost of their implementations. then propose high-radix algorithm aimed at short execution time low cost. present detailed radix-8 based 256-bit primes Verilog HDL compare its with other Our Altera Cyclone V FPGA chip uses...

Numerous recent approaches to modeling and re-rendering dynamic scenes leverage plane-based explicit representations, addressing slow training times associated with models like neural radiance fields (NeRF) Gaussian splatting (GS). However, merely decomposing 4D into multiple 2D representations is insufficient for high-fidelity of complex motions. In response, we present DaRePlane, a novel direction-aware representation approach that captures scene dynamics from six different directions....

The Segment Anything Model 2 (SAM 2) is the latest generation foundation model for image and video segmentation. Trained on expansive Video (SA-V) dataset, which comprises 35.5 million masks across 50.9K videos, SAM advances its predecessor's capabilities by supporting zero-shot segmentation through various prompts (e.g., points, boxes, masks). Its robust performance efficient memory usage make particularly appealing surgical tool in especially given scarcity of labeled data diversity...

We present a theoretical study of the H+OCl system on an accurate ab initio potential energy surface (PES) investigated by Peterson et al.[J. Chem. Phys. 113 (2000) 6186]. Both exact time-dependent quantum wave packet (TDWP) and quasi-classical trajectory (QCT) methods are employed. The results reaction probabilities for total angular momentum J = 0 integral cross section calculated TDWP in good agreement with QCT ones. Additionally, nearly forward-backward symmetric product scattering...