Radhakrishnan Gopalapillai
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Radhakrishnan Gopalapillai has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
- Deepa Gupta
Amrita School of Computing, Amrita Vishwa Vidyapeetham
10 shared publications - Tekal Subramanyam Babu Sudarshan
PES University
8 shared publications - Priyanka C. Nair
Amrita School of Engineering, Bengaluru, Amrita Vishwa Vidyapeetham, Amrita University
2 shared publications - Amudha Joseph
Amrita Vishwa Vidyapeetham
1 shared publication Ankita Ajith
School of Engineering, Amrita Vishwa Vidyapeetham, Bangalore Campus, India
1 shared publicationDeepa Gupta
1 shared publication
Latest Publications
- 1 May 2026
- Published by Springer Nature
- p. 35-46
Cellular Automata for Urban Growth Predictions: Review, Simulation, and Prospects
- 1 May 2026
- Published by Springer Nature
- p. 35-46
Cellular Automata for Urban Growth Predictions: Review, Simulation, and Prospects
- 19 May 2023
- Published by Springer Nature
- p. 53-63
Prediction of COVID-19 Pandemic Spread Using Graph Neural Networks
- 19 May 2023
- Published by Springer Nature
- p. 53-63
Prediction of COVID-19 Pandemic Spread Using Graph Neural Networks
- Open Access
Convolution-Based Encoding of Depth Images for Transfer Learning in RGB-D Scene Classification
Classification of indoor environments is a challenging problem. The availability of low-cost depth sensors has opened up a new research area of using depth information in addition to color image (RGB) data for scene understanding. Transfer learning of deep convolutional networks with pairs of RGB and depth (RGB-D) images has to deal with integrating these two modalities. Single-channel depth images are often converted to three-channel images by extracting horizontal disparity, height above ground, and the angle of the pixel’s local surface normal (HHA) to apply transfer learning using networks trained on the Places365 dataset. The high computational cost of HHA encoding can be a major disadvantage for the real-time prediction of scenes, although this may be less important during the training phase. We propose a new, computationally efficient encoding method that can be integrated with any convolutional neural network. We show that our encoding approach performs equally well or better in a multimodal transfer learning setup for scene classification. Our encoding is implemented in a customized and pretrained VGG16 Net. We address the class imbalance problem seen in the image dataset using a method based on the synthetic minority oversampling technique (SMOTE) at the feature level. With appropriate image augmentation and fine-tuning, our network achieves scene classification accuracy comparable to that of other state-of-the-art architectures.
Open AccessConvolution-Based Encoding of Depth Images for Transfer Learning in RGB-D Scene Classification
Classification of indoor environments is a challenging problem. The availability of low-cost depth sensors has opened up a new research area of using depth information in addition to color image (RGB) data for scene understanding. Transfer learning of deep convolutional networks with pairs of RGB and depth (RGB-D) images has to deal with integrating these two modalities. Single-channel depth images are often converted to three-channel images by extracting horizontal disparity, height above ground, and the angle of the pixel’s local surface normal (HHA) to apply transfer learning using networks trained on the Places365 dataset. The high computational cost of HHA encoding can be a major disadvantage for the real-time prediction of scenes, although this may be less important during the training phase. We propose a new, computationally efficient encoding method that can be integrated with any convolutional neural network. We show that our encoding approach performs equally well or better in a multimodal transfer learning setup for scene classification. Our encoding is implemented in a customized and pretrained VGG16 Net. We address the class imbalance problem seen in the image dataset using a method based on the synthetic minority oversampling technique (SMOTE) at the feature level. With appropriate image augmentation and fine-tuning, our network achieves scene classification accuracy comparable to that of other state-of-the-art architectures.
- 24 April 2019
- Published by Springer Nature
- p. 997-1010
Classification of Service Robot Environments Using Multimodal Sequence Data
- 24 April 2019
- Published by Springer Nature
- p. 997-1010
Classification of Service Robot Environments Using Multimodal Sequence Data
- Open Access
- 1 January 2017
- Published by Elsevier in Procedia Computer Science
- Vol. 115, 63-71
Pattern Identification of Robotic Environments using Machine Learning Techniques
- 1 January 2017
- Published by Elsevier in Procedia Computer Science
- Vol. 115, 63-71
Open AccessPattern Identification of Robotic Environments using Machine Learning Techniques
- 25 December 2015
- Published by Springer Nature
- p. 521-531
Multi-view Robotic Time Series Data Clustering and Analysis Using Data Mining Techniques
- 25 December 2015
- Published by Springer Nature
- p. 521-531
Multi-view Robotic Time Series Data Clustering and Analysis Using Data Mining Techniques
- 1 November 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 208-213
Clustering of robotic environment using image data stream
Mobile robots are being used in various applications like space shuttles, intelligent home security, military applications or other service oriented applications where human intervention is limited. A robot has to understand its environment by analyzing the data to take the appropriate actions in the given environment. Mostly the data collected from the sensors on the robots are huge and continuous, making it impossible to store the entire data in main memory and hence allowing only single scan of data. Traditional clustering algorithms like k-means cannot be used in such environment as they require multiple scan of data. This paper presents an experimental study on the implementation of Stream KM++, a data stream clustering algorithm that effectively cluster these time series robotic image data within the memory restrictions under various conditions. Promising results are obtained from the various experiments carried out.
- 1 November 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 208-213
Clustering of robotic environment using image data stream
Mobile robots are being used in various applications like space shuttles, intelligent home security, military applications or other service oriented applications where human intervention is limited. A robot has to understand its environment by analyzing the data to take the appropriate actions in the given environment. Mostly the data collected from the sensors on the robots are huge and continuous, making it impossible to store the entire data in main memory and hence allowing only single scan of data. Traditional clustering algorithms like k-means cannot be used in such environment as they require multiple scan of data. This paper presents an experimental study on the implementation of Stream KM++, a data stream clustering algorithm that effectively cluster these time series robotic image data within the memory restrictions under various conditions. Promising results are obtained from the various experiments carried out.
- 1 July 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Experimentation and analysis of time series data from multi-path robotic environment
Autonomous mobile robots are increasingly used in many application areas. In most applications, they have to explore and gather knowledge about the environment they are deployed in. These robots transfer real time data about the environment continuously. This paper discusses a set of experiments that have been carried out to simulate various robotic environments. A robot attached with four sensors is used to collect information about the environment as the robot moves in multiple straight line paths. Time series data collected from these experiments are clustered using data mining techniques. Experimental results show clustering accuracies vary depending on the number of clusters formed.
- 1 July 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Experimentation and analysis of time series data from multi-path robotic environment
Autonomous mobile robots are increasingly used in many application areas. In most applications, they have to explore and gather knowledge about the environment they are deployed in. These robots transfer real time data about the environment continuously. This paper discusses a set of experiments that have been carried out to simulate various robotic environments. A robot attached with four sensors is used to collect information about the environment as the robot moves in multiple straight line paths. Time series data collected from these experiments are clustered using data mining techniques. Experimental results show clustering accuracies vary depending on the number of clusters formed.
- 12 December 2014
- Published by Springer Nature
- p. 489-498
Acquisition and Analysis of Robotic Data Using Machine Learning Techniques
- 12 December 2014
- Published by Springer Nature
- p. 489-498
Acquisition and Analysis of Robotic Data Using Machine Learning Techniques