Multi-Instance Convolutional Neural Network for multi-shot person re-identification

Liu, Xiaokai; Bi, Sheng; Ma, Xiaorui; Wang, Jie

doi:10.1016/j.neucom.2019.01.076

Cited by 8 publications

(3 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MIL can be utilized for tasks like activity recognition or identifying anomalies in video surveillance [61][62][63]. Bags represent video clips or sequences, while instances are specific time frames inside such videos.…”

Section: Video Classificationmentioning

confidence: 99%

A Comprehensive Review on Multiple Instance Learning

Fatima,

Ali,

Kim

2023

Electronics

View full text Add to dashboard Cite

Multiple-instance learning has become popular over recent years due to its use in some special scenarios. It is basically a type of weakly supervised learning where the learning dataset contains bags of instances instead of a single feature vector. Each bag is associated with a single label. This type of learning is flexible and a natural fit for multiple real-world problems. MIL has been employed to deal with a number of challenges, including object detection and identification tasks, content-based image retrieval, and computer-aided diagnosis. Medical image analysis and drug activity prediction have been the main uses of MIL in biomedical research. Many Algorithms based on MIL have been put forth over the years. In this paper, we will discuss MIL, the background of MIL and its application in multiple domains, some MIL-based methods, challenges, and lastly, the conclusions and prospects.

show abstract

Section: Video Classificationmentioning

confidence: 99%

A Comprehensive Review on Multiple Instance Learning

Fatima,

Ali,

Kim

2023

Electronics

View full text Add to dashboard Cite

show abstract

“…This Paper [11] described an instance-level recognition problem where a person re-id relying on the discriminative features capturing not only the different spatial scales but also incorporating an arbitrary combination of multiple scales. A novel deep re-id neural network was designed and was termed as Omni-Scale network (Osnet) to achieve a residual block composed of multiple convolutional streams which can detect features at a certain scale.…”

Section: Related Workmentioning

confidence: 99%

Person Re-Identification Using Deep Learning Approach: A Survey

Pandeeswari¹,

Shargunam²,

Rajakumar³

2021

AJEAT

View full text Add to dashboard Cite

Person Re-identification proof crosswise over various observation cameras with disjoint fields of view has turned out to be one of the most intriguing and testing subjects in the zone of wise video observation. Albeit a few techniques have been created, what's more, proposed, certain confinements and uncertain issues remain. In the majority of the current Re-identification proof approaches, highlight vectors are separated from sectioned still pictures or video outlines. Distinctive similitude or disparity measures have been applied to these vectors. A few strategies have utilized basic consistent measurements, while others have used models to acquire streamlined measurements. Some have made models dependent on neighborhood shading or surface data, and others have constructed models dependent on the stride of individuals. When all is said in done, the primary goal of every one of these methodologies is to accomplish a higher-exactness rate and lower computational costs.

show abstract

“…Instance re-identification matches image patches for object re-identification purposes. The most similar problem to our task is the person re-identification problem, which has become a major research interest recently [8,[29][30][31][32][33][34]. Another interesting application includes vehicle re-identification.…”

Section: Related Workmentioning

confidence: 99%

Multi-View Instance Matching with Learned Geometric Soft-Constraints

Nassar

Lefèvre

Wegner

2020

IJGI

View full text Add to dashboard Cite

We present a new approach for matching urban object instances across multiple ground-level images for the ultimate goal of city-scale mapping of objects with high positioning accuracy. What makes this task challenging is the strong change in view-point, different lighting conditions, high similarity of neighboring objects, and variability in scale. We propose to turn object instance matching into a learning task, where image-appearance and geometric relationships between views fruitfully interact. Our approach constructs a Siamese convolutional neural network that learns to match two views of the same object given many candidate image cut-outs. In addition to image features, we propose utilizing location information about the camera and the object to support image evidence via soft geometric constraints. Our method is compared to existing patch matching methods to prove its edge over state-of-the-art. This takes us one step closer to the ultimate goal of city-wide object mapping from street-level imagery to benefit city administration.

show abstract

Multi-Instance Convolutional Neural Network for multi-shot person re-identification

Cited by 8 publications

References 9 publications

A Comprehensive Review on Multiple Instance Learning

A Comprehensive Review on Multiple Instance Learning

Person Re-Identification Using Deep Learning Approach: A Survey

Multi-View Instance Matching with Learned Geometric Soft-Constraints

Contact Info

Product

Resources

About