Analysis of Model based Shadow Detection and Removal in Color Images

Ghewari, Trupti Rajendra; Khot, S. R.; Khatavkar, Mandar Digambar

doi:10.1109/icisc44355.2019.9036408

Cited by 2 publications

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This section explores the recent related work considering main focus on the approach, applicability, challenges and results. In 2019, Ghewari et al [25] presented a successive thresholding scheme (STS) which was taken from the Tsai's method using the global thresholding scheme.…”

Section: Literature Reviewmentioning

confidence: 99%

Clustering based ACO and ABC algorithms for the shadow detection and removal

2022

IJATEE

View full text Add to dashboard Cite

Shadow detection and removal is challenging due to different factors like image context, light conditions, complexity etc. [1-3]. So, accruing, detection, convergence and removal in terms of computer vision is a challenging task [4-6]. There are several algorithms were explored in the same direction for the proper detection and removal. Clustering, classification and nature inspired algorithms were used mostly [7, 8]. Main clustering algorithms are kmeans, fuzzy c-means (FCM) and hierarchical clustering. These algorithms are beneficial in key grouping and shadow pointing [9-12]. Ant colony optimization (ACO), particle swarm optimization (PSO), teaching learning-based optimization (TLBO), cuckoo search, artificial bee colony (ABC), etc. are the main nature inspired algorithms [13,14]. *Author for correspondence Classification algorithms mainly used for the same are decision tree (DT), random forest (RF), support vector machine (SVM), logistic regression (LR), naïve Bayes (NB), k-nearest neighbor (KNN), etc. [15-24].The following problems were identified during the time of the review and analysis. This was annotated based on the analysis and the observations of the complete study. It shows the analytical and explorative way of exploration of the purpose and the study scenarios. 1. There is the need of exploration of surface complexity as it may affect the detection results. 2. There is the need to cover multiple correlated attributes like intensity, shape, brightness, complexity etc. in terms of the overall analysis and complete detection computation. 3. There is the need of clustering mechanism for the selection of key frames for different lighting and ecological conditions. It can be further segmented for better accuracy.

show abstract

Section: Literature Reviewmentioning

confidence: 99%