(Press-News.org)
Vehicle Color Recognition (VCR) is vital in intelligent traffic management and criminal investigation assistance. However, the existing vehicle color datasets only cover 13 classes, which can not meet the current actual demand. Besides, although lots of efforts are devoted to VCR, they suffer from the problem of class imbalance in datasets.
To solve the problems, a research team led by Mingdi HU published their new research in Frontiers of Computer Science co-published by Higher Education Press and Springer·Nature.
The team propose a novel VCR method based on Smooth Modulation Neural Network with Multi-Scale Feature Fusion (SMNN-MSFF). They present a new VCR dataset with 24 vehicle classes, Vehicle Color-24. And propose the SMNN-MSFF model with multiscale feature fusion and smooth modulation. The former aims to extract feature information from local to global, and the latter could increase the loss of the images of tail class instances for training with class-imbalance. Extensive ablation studies demonstrate that each module of the proposed method is effective, especially the smooth modulation efficiently helps feature learning of the minority or tail classes. Comprehensive experimental evaluation on Vehicle Color-24 and previously three representative datasets demonstrate that the proposed SMNN-MSFF outperformed state-of-the-art VCR methods.
In the research, they built a new dataset with 24 vehicle colors, called Vehicle Color-24. The colors of Vehicle Color-24 are divided into 24 types, including red, dark-red, pink, orange, dark-orange, red-orange, yellow, lemon-yellow, earthy-yellow, green, dark-green, grass-green, cyan, blue, dark-blue, purple, black, white, silver-gray, gray, dark-gray, champagne, brown and dark-brown. Vehicle Color-24 can make up for the current needs of practical vehicle traffic management and criminal vehicle tracking applications.
Then, they propose a novel vehicle color recognition method based on SMNN-MSFF. Firstly, this algorithm starts to pay attention to the color distribution imbalance nature existing in any dataset. The loss function fine-tunes the network so that the algorithm can better capture the characteristics of small-scale classes than focal loss through ablation experiments. Secondly, this network adds an FPN module to extract edges and corners information, which is helpful to extract vehicle shape features and local location information to assist vehicle recognition. Thirdly, this backbone network is designed with only 42 layers, which belong to a lightweight network, to relieve the pressure of storage and increase the possibility of implementation in practical applications.
The experimental results show that the mAP of our method in our paper is 94.96% in recognizing 24 types of colors. The proposed SMNN-MSFF outperformed state-of-the-art VCR methods, and better meet the requirements for fine classification of vehicle colors.
However, since the actual environment can be affected by unpredictable factors and the long tail effect exists in vehicle color distribution, further efforts to improve the fine recognition of vehicle color are still required. Future work will continue studying the solution of class imbalance because the vehicle color is diverse, and the vehicle color dataset must have the characteristics of the long-tail distribution.
###
Research Article, Published: 15 June 2023
Mingdi HU, Long BAI, Jiulun FAN, Sirui ZHAO, Enhong CHEN. Vehicle color recognition based on smooth modulation neural network with multi-scale feature fusion. Front. Comput. Sci., 2023, 17(3): 173321,
https://doi.org/10.1007/s11704-022-1389-x
About Frontiers of Computer Science (FCS)
FCS was launched in 2007. It is published bimonthly both online and in print by HEP and Springer. Prof. Zhi-Hua Zhou from Nanjing University serves as the Editor-in-Chief. It aims to provide a forum for the publication of peer-reviewed papers to promote rapid communication and exchange between computer scientists. FCS covers all major branches of computer science, including: architecture, software, artificial intelligence, theoretical computer science, networks and communication, information systems, multimedia and graphics, information security, interdisciplinary, etc. The readers may be interested in the special columns "Perspective" and "Excellent Young Scholars Forum".
FCS is indexed by SCI(E), EI, DBLP, Scopus, etc. The latest IF is 2.669. FCS solicits the following article types: Review, Research Article, Letter.
END
The Science Partner Journal (SPJ) program is pleased to announce the inclusion of five SPJs in the 2023 Journal Citation Reports (JCR) from Clarivate.
Research, launched in 2018, received its second Journal Impact Factor (JIF) this year, 11.0, ranking it #9 among 73 journals in the Multidisciplinary category in Science Citation Index Expanded (SCIE).
Plant Phenomics also launched in 2018 and received its second JIF this year, 6.5, ranking it #5 among 88 journals in the Agronomy, #21 among ...
Scientists identify the first genetic marker for MS severity, opening the door to preventing long-term disability.
A study of more than 22,000 people with multiple sclerosis has discovered the first genetic variant associated with faster disease progression that can rob patients of their mobility and independence over time. Multiple sclerosis (MS) is the result of the immune system mistakenly attacking the brain and the spinal cord, resulting in symptom flares known as relapses as well as longer-term degeneration known as progression. Despite the development of effective treatments for relapses, ...
ITHACA, N.Y. -- When it comes to e-cigarette warning labels, respondents in focus groups organized by Cornell researchers were clear: Give it to me straight.
But approximately 20 years after they hit the market, electronic cigarettes’ precise health risks remain unclear. And for adults trying to quit smoking conventional cigarettes, ambiguity in messaging can skew perceptions of the health benefits of using these products as an alternative to combustible cigarettes.
A multidisciplinary team led by Jeff ...
An ingredient in many toothpastes is sodium fluoride, a compound of fluorine. It is added to protect teeth against decay. But compounds containing fluorine have other practical uses that might surprise you. Scientists at the U.S. Department of Energy’s (DOE) Argonne National Laboratory have discovered a fluoride electrolyte that could protect a next generation battery against performance decline.
“An exciting new generation of battery types for electric vehicles beyond lithium ion is on the horizon,” said Zhengcheng (John) Zhang, a group leader in Argonne’s Chemical Sciences and Engineering division.
The ...
Male birds that are able to repeat song notes precisely stand the best chance of attracting a female mate, according to a new study published in Nature Communications.
However, the males need to ensure they have a selection of different songs in their repertoire if they are to hold a female’s attention and prevent her from getting bored.
The findings from the study, by scientists from Lancaster University and Manchester Metropolitan University, sheds new light on the evolution of bird song.
Over ...
A savvy hacker can obtain secret information, such as a password, by observing a computer program’s behavior, like how much time that program spends accessing the computer’s memory.
Security approaches that completely block these “side-channel attacks” are so computationally expensive that they aren’t feasible for many real-world systems. Instead, engineers often apply what are known as obfuscation schemes that seek to limit, but not eliminate, an attacker’s ability to learn secret information.
To ...
A researcher at the Advanced Manufacturing Institute and the Texas Center for Superconductivity at the University of Houston (TCSUH) has found a way to reduce superconductor failures, enabled by a Pulsed Laser Deposition (PLD) tool. The popular thin film deposition instrument will be purchased with an $800,000 grant from the U.S. Office of Naval Research.
At extremely low temperatures (as low as cryogenic temperatures), superconductors allow electric current to flow without resistance and produce strong magnetic fields. That’s the principle behind Magnetic Resonance Imaging (MRI) ...
A new technique developed by UCLA geologists that uses artificial intelligence to better predict where and why landslides may occur could bolster efforts to protect lives and property in some of the world’s most disaster-prone areas.
The new method, described in a paper published in the journal Communications Earth & Environment, improves the accuracy and interpretability of AI-based machine-learning techniques, requires far less computing power and is more broadly applicable than traditional predictive models.
The approach would be particularly valuable in places like California, the researchers say, where drought, ...
In June, the American Medical Association announced a new policy that encourages physicians not to focus solely on body mass index, or BMI, as a determinant of weight and health. As a ratio of an individual’s weight and height, BMI can provide an easy and inexpensive but often misleading measure of someone’s overall health.
While this new policy is not mandatory for physicians, it is part of a growing opinion that BMI is more useful for assessing population health rather than individual ...
As battlespace sensors proliferate and data increases, commanders can easily find themselves in an information paradox: drowning in data, but starving for knowledge.
U.S. Marine Corps Lt. Col. Pedro Ortiz, who graduated from the Naval Postgraduate School (NPS) on June 16 with a Ph.D. in Computer Science, focused his dissertation on this challenge to help enable rapid, effective decision-making for commanders in an era of ever-increasing sensor data and uncertainty.
“I am very interested in applying artificial intelligence and machine learning to solve warfighter problems,” ...
