Thanikaiselvan Veeramalai
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Thanikaiselvan Veeramalai has a total of 92 co-authors. Here are the top co-authors by number of shared publications:
- Rengarajan Amirtharajan
School of Electrical & Electronics Engineering, SASTRA Deemed University, Thanjavur, India
12 shared publications - Pachaiyappan Arulmozhivarman
VIT University
6 shared publications - Ahmad Shaik
Vellore Institute of Technology
5 shared publications - Shounak Shastri
VIT University
5 shared publications - John Bosco Balaguru Rayappan
Centre for Nanotechnology & Advanced Biomaterials (CeNTAB), SASTRA Deemed University, Thanjavur, India
4 shared publications - Sakshi Patel
School of Electronics Engineering, Vellore Institute of Technology (VIT), Vellore, India
3 shared publications
Latest Publications
- Open Access
- 27 October 2026
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 9, 1919612
3DSLCM-CNN: 3D sinusoidal logistic chaotic map and CNN for image encryption
A new approach for image encryption is introduced in this article with a deep network guided key map/image for diffusion. The limitations in the selection or generation of key images for diffusion in the existing chaotic map-based methods are focused on in this work. The sinusoidal logistic map is modified and extended to three dimensions in the proposed work for confusion. A variant of the sinusoidal logistic function is utilized to generate a random sequence and resized to the size of 2 m × 2 n and normalized to the range of pixel intensity of plaintext image having size m × n . The size of key map is reduced equal to the size of plaintext image using a deep network with dilated convolution. The plaintext image is scrambled using row-column shuffling and split into two image parts, odd and even images. Both the images are then reshaped and then confused using the proposed chaotic map and then combined into a single image. The confused image is diffused with the deep network-guided key map at the bit level using XOR logic operation. To enhance the security, the diffused image is further confused using the chaotic maps. The overall process is reversed in decryption to retrieve the plaintext image. The performance of the proposed encryption is measured and compared with the existing methods using standard measures such as key sensitivity, information entropy, correlation coefficient, NPCR, and UACI, which shows that the proposed method improves randomness to secure the information from attacks.
- 27 October 2026
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 9, 1919612
Open Access3DSLCM-CNN: 3D sinusoidal logistic chaotic map and CNN for image encryption
A new approach for image encryption is introduced in this article with a deep network guided key map/image for diffusion. The limitations in the selection or generation of key images for diffusion in the existing chaotic map-based methods are focused on in this work. The sinusoidal logistic map is modified and extended to three dimensions in the proposed work for confusion. A variant of the sinusoidal logistic function is utilized to generate a random sequence and resized to the size of 2 m × 2 n and normalized to the range of pixel intensity of plaintext image having size m × n . The size of key map is reduced equal to the size of plaintext image using a deep network with dilated convolution. The plaintext image is scrambled using row-column shuffling and split into two image parts, odd and even images. Both the images are then reshaped and then confused using the proposed chaotic map and then combined into a single image. The confused image is diffused with the deep network-guided key map at the bit level using XOR logic operation. To enhance the security, the diffused image is further confused using the chaotic maps. The overall process is reversed in decryption to retrieve the plaintext image. The performance of the proposed encryption is measured and compared with the existing methods using standard measures such as key sensitivity, information entropy, correlation coefficient, NPCR, and UACI, which shows that the proposed method improves randomness to secure the information from attacks.
- Open Access
- 14 October 2026
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 9, 1913975
Deep network-guided key map generation for lossless image encryption
In this article, a novel 4D Non-monotonic Sinusoidal Chaotic Map (NSCM) for confusion and a deep network-guided key map generation for diffusion are introduced to improve information security against various attacks. This article utilizes the advantages of the sinusoidal function to modify the monotonic function for generating random sequences to obtain a chaotic map for permutation. The limitations of substitution, such as chosen/known plaintext attacks, are suppressed using deep network-based key map generation. A key map equal in size to the size of the plaintext image is obtained as a generative output from a deep network using a smaller map and normalized to the intensity range of the plaintext image. The plaintext image is scrambled block-wise, followed by pixel-wise confusion using the chaotic map. Then the scrambled image is diffused with the key map using a bit-level XOR operation to obtain the encrypted image. The proposed method can be reversed for retrieving the plaintext image, that is, diffusion followed by confusion. The experimental results show that the proposed method performs better qualitatively in terms of visual perception as well as quantitatively in terms of entropy and Peak Signal-to-Noise Ratio (PSNR). Also, the proposed method strongly restricts various attacks such as chosen-plaintext attacks and differential attacks.
- 14 October 2026
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 9, 1913975
Open AccessDeep network-guided key map generation for lossless image encryption
In this article, a novel 4D Non-monotonic Sinusoidal Chaotic Map (NSCM) for confusion and a deep network-guided key map generation for diffusion are introduced to improve information security against various attacks. This article utilizes the advantages of the sinusoidal function to modify the monotonic function for generating random sequences to obtain a chaotic map for permutation. The limitations of substitution, such as chosen/known plaintext attacks, are suppressed using deep network-based key map generation. A key map equal in size to the size of the plaintext image is obtained as a generative output from a deep network using a smaller map and normalized to the intensity range of the plaintext image. The plaintext image is scrambled block-wise, followed by pixel-wise confusion using the chaotic map. Then the scrambled image is diffused with the key map using a bit-level XOR operation to obtain the encrypted image. The proposed method can be reversed for retrieving the plaintext image, that is, diffusion followed by confusion. The experimental results show that the proposed method performs better qualitatively in terms of visual perception as well as quantitatively in terms of entropy and Peak Signal-to-Noise Ratio (PSNR). Also, the proposed method strongly restricts various attacks such as chosen-plaintext attacks and differential attacks.
- Open Access
- 6 January 2026
- Published by Elsevier in Results in Engineering
- Vol. 29
Autoencoder-based image encryption using hybrid scrambling, diffusion, and dimensionality reduction
- 6 January 2026
- Published by Elsevier in Results in Engineering
- Vol. 29
Open AccessAutoencoder-based image encryption using hybrid scrambling, diffusion, and dimensionality reduction
- Open Access
- 6 December 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1)
Image adaptive encryption using EfficientNet B3 feature guided multi scroll chaotic map with modulo controlled pseudo parallel processing
A new multistage encryption algorithm is proposed by integrating the deep neural network with a new 4D multi-scroll chaotic map to enhance the efficiency and improve the security of image transmission in the open channel. This combined network expands the key space and maintains the secrecy of the key with the multistage encryption algorithm. Initially, the image adaptive key generation process is implemented by the EfficientNet-B3 network to extract the features from the source image, which are then converted into hash values using SHA 256. The hash values are partitioned into four sections, and each section is normalized to give one distinct initial value for the generation of a multi-scroll chaotic sequence. The pseudo-parallel process routes the split source sub-image blocks (128⨯128) of plain text to branch 1 or branch 2, decided by the seed value of the chaotic sequence, increasing the high robustness against the differential and statistical attacks. Each branch contains row and column-wise permutations, bidirectional selective shuffling, and chaotic intra/inter-pixel diffusion in varying orders. The key image diffusion and dynamic DNA diffusion to the intermediate cypher image exhibit a strong avalanche effect. The simulation evaluation on the natural data set images demonstrates the large key space of 2 to the power of 674, high key sensitivity, uniform histogram with entropy value attains the critical values of 7.9, high NPCR value of 99.9%, UACI values with 33.46%, almost zero-pixel correlation and strong robustness to the cropping and noise attacks.
- 6 December 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1)
Open AccessImage adaptive encryption using EfficientNet B3 feature guided multi scroll chaotic map with modulo controlled pseudo parallel processing
A new multistage encryption algorithm is proposed by integrating the deep neural network with a new 4D multi-scroll chaotic map to enhance the efficiency and improve the security of image transmission in the open channel. This combined network expands the key space and maintains the secrecy of the key with the multistage encryption algorithm. Initially, the image adaptive key generation process is implemented by the EfficientNet-B3 network to extract the features from the source image, which are then converted into hash values using SHA 256. The hash values are partitioned into four sections, and each section is normalized to give one distinct initial value for the generation of a multi-scroll chaotic sequence. The pseudo-parallel process routes the split source sub-image blocks (128⨯128) of plain text to branch 1 or branch 2, decided by the seed value of the chaotic sequence, increasing the high robustness against the differential and statistical attacks. Each branch contains row and column-wise permutations, bidirectional selective shuffling, and chaotic intra/inter-pixel diffusion in varying orders. The key image diffusion and dynamic DNA diffusion to the intermediate cypher image exhibit a strong avalanche effect. The simulation evaluation on the natural data set images demonstrates the large key space of 2 to the power of 674, high key sensitivity, uniform histogram with entropy value attains the critical values of 7.9, high NPCR value of 99.9%, UACI values with 33.46%, almost zero-pixel correlation and strong robustness to the cropping and noise attacks.
- Open Access
- 2 July 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-31
Enhanced security for medical images using a new 5D hyper chaotic map and deep learning based segmentation
Medical image encryption is important for maintaining the confidentiality of sensitive medical data and protecting patient privacy. Contemporary healthcare systems store significant patient data in text and graphic form. This research proposes a New 5D hyperchaotic system combined with a customised U-Net architecture. Chaotic maps have become an increasingly popular method for encryption because of their remarkable characteristics, including statistical randomness and sensitivity to initial conditions. The significant region is segmented from the medical images using the U-Net network, and its statistics are utilised as initial conditions to generate the new random sequence. Initially, zig-zag scrambling confuses the pixel position of a medical image and applies further permutation with a new 5D hyperchaotic sequence. Two stages of diffusion are used, such as dynamic DNA flip and dynamic DNA XOR, to enhance the encryption algorithm’s security against various attacks. The randomness of the New 5D hyperchaotic system is verified using the NIST SP800-22 statistical test, calculating the Lyapunov exponent and plotting the attractor diagram of the chaotic sequence. The algorithm validates with statistical measures such as PSNR, MSE, NPCR, UACI, entropy, and Chi-square values. Evaluation is performed for test images yields average horizontal, vertical, and diagonal correlation coefficients of –0.0018, –0.0002, and 0.0007, respectively, Shannon entropy of 7.9971, Kolmogorov Entropy value of 2.9469, NPCR of 99.61%, UACI of 33.49%, Chi-square “PASS” at both the 5% (293.2478) and 1% (310.4574) significance levels, key space is 2500 and an average encryption time of approximately 2.93 s per 256 × 256 image on a standard desktop CPU. The performance comparisons use various encryption methods and demonstrate that the proposed method ensures secure reliability against various challenges.
- 2 July 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-31
Open AccessEnhanced security for medical images using a new 5D hyper chaotic map and deep learning based segmentation
Medical image encryption is important for maintaining the confidentiality of sensitive medical data and protecting patient privacy. Contemporary healthcare systems store significant patient data in text and graphic form. This research proposes a New 5D hyperchaotic system combined with a customised U-Net architecture. Chaotic maps have become an increasingly popular method for encryption because of their remarkable characteristics, including statistical randomness and sensitivity to initial conditions. The significant region is segmented from the medical images using the U-Net network, and its statistics are utilised as initial conditions to generate the new random sequence. Initially, zig-zag scrambling confuses the pixel position of a medical image and applies further permutation with a new 5D hyperchaotic sequence. Two stages of diffusion are used, such as dynamic DNA flip and dynamic DNA XOR, to enhance the encryption algorithm’s security against various attacks. The randomness of the New 5D hyperchaotic system is verified using the NIST SP800-22 statistical test, calculating the Lyapunov exponent and plotting the attractor diagram of the chaotic sequence. The algorithm validates with statistical measures such as PSNR, MSE, NPCR, UACI, entropy, and Chi-square values. Evaluation is performed for test images yields average horizontal, vertical, and diagonal correlation coefficients of –0.0018, –0.0002, and 0.0007, respectively, Shannon entropy of 7.9971, Kolmogorov Entropy value of 2.9469, NPCR of 99.61%, UACI of 33.49%, Chi-square “PASS” at both the 5% (293.2478) and 1% (310.4574) significance levels, key space is 2500 and an average encryption time of approximately 2.93 s per 256 × 256 image on a standard desktop CPU. The performance comparisons use various encryption methods and demonstrate that the proposed method ensures secure reliability against various challenges.
- Open Access
- 5 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 131928-131941
FibroRegNet: A Regression Framework for the Pulmonary Fibrosis Prognosis Prediction Using a Convolutional Spatial Transformer Network
Predicting the growth of idiopathic pulmonary fibrosis (IPF) is crucial for effectively treating patients affected by the disease. While the Forced Vital Capacity (FVC) serves as one of the indicators of lung functionality, accurately determining its decline solely based on previous FVC values presents a significant obstacle. We propose the utilization of a multimodal system called FibroRegNet, which capitalizes on the recent achievements in cross-model learning across general domains. FibroRegNet is designed to acquire knowledge through the regression function which maps the multimodal inputs, including CT scan and demographic information, to the coefficients of the quadratic polynomial ridge regression of FVC as outputs. FibroRegNet estimates the lung volume from a fraction of CT slices, encodes the demographic information, and combines these features with the convolutional features, from selected CT slices, that are learned through convolutional spatial transformer modules in three identical parallel streams. Trained on a publicly available database, FibroRegNet has shown significant improvement in the results compared to the related past works with a modified Laplace log-likelihood score of −6.64. Furthermore, we believe that this network has the potential to provide advantages in research domains related to the development of networks aimed at enhancing the predictive precision of IPF.
- 5 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 131928-131941
Open AccessFibroRegNet: A Regression Framework for the Pulmonary Fibrosis Prognosis Prediction Using a Convolutional Spatial Transformer Network
Predicting the growth of idiopathic pulmonary fibrosis (IPF) is crucial for effectively treating patients affected by the disease. While the Forced Vital Capacity (FVC) serves as one of the indicators of lung functionality, accurately determining its decline solely based on previous FVC values presents a significant obstacle. We propose the utilization of a multimodal system called FibroRegNet, which capitalizes on the recent achievements in cross-model learning across general domains. FibroRegNet is designed to acquire knowledge through the regression function which maps the multimodal inputs, including CT scan and demographic information, to the coefficients of the quadratic polynomial ridge regression of FVC as outputs. FibroRegNet estimates the lung volume from a fraction of CT slices, encodes the demographic information, and combines these features with the convolutional features, from selected CT slices, that are learned through convolutional spatial transformer modules in three identical parallel streams. Trained on a publicly available database, FibroRegNet has shown significant improvement in the results compared to the related past works with a modified Laplace log-likelihood score of −6.64. Furthermore, we believe that this network has the potential to provide advantages in research domains related to the development of networks aimed at enhancing the predictive precision of IPF.
- Open Access
- 29 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 107059-107076
Balancing of Attenuation Disparity to Restore the Weak Color Channels in Underwater Images
The images captured under water are typically deteriorated due to non uniform attenuation of traversing light in underwater medium. The studies on underwater environment mainly depends on the techniques for the enhancement of underwater images. Many state-of-the-arts compensate the losses due to scattering of light in underwater images but still it is a challenging task to compensate the color distortion due to absorption of light. A new approach is introduced in this article for the enhancement of underwater images by balancing the attenuation disparity between the color channels. This approach reduces the color dominancy based on the prior that the non-uniform attenuation of RGB channels of light in water which results in severely absorbed, moderately absorbed and informative color channels. The difference between the maximum pixel intensity level of an image and its color channels is considered as the attenuation disparity of color channels. So, this leads to the statistical classification of underwater images into five major types: haze, bluish, greenish, greenish blue and bluish green images and two special types such as yellowish image and dark image subjectively. The pixel intensity of severely and moderately absorbed color channels are manipulated in accordance with the informative color channel by adding the value of concerned attenuation disparity globally. It increases the pixel intensity level of weak color channels and restores the channel information but does not improve the visual quality. The Gray World (GW) method or color equalization (CE) method is applied to remove the color imbalance and the contrast is improved by saturating the high level and low level pixel intensities. Finally, the conventional multiscale image fusion is employed to improve the image quality. The experimental results show that the proposed method outperforms qualitatively and quantitatively in terms of underwater color image quality evaluation (UCIQE) metrics and underwater image quality measures (UIQM) compared to the existing image enhancement methods.
- 29 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 107059-107076
Open AccessBalancing of Attenuation Disparity to Restore the Weak Color Channels in Underwater Images
The images captured under water are typically deteriorated due to non uniform attenuation of traversing light in underwater medium. The studies on underwater environment mainly depends on the techniques for the enhancement of underwater images. Many state-of-the-arts compensate the losses due to scattering of light in underwater images but still it is a challenging task to compensate the color distortion due to absorption of light. A new approach is introduced in this article for the enhancement of underwater images by balancing the attenuation disparity between the color channels. This approach reduces the color dominancy based on the prior that the non-uniform attenuation of RGB channels of light in water which results in severely absorbed, moderately absorbed and informative color channels. The difference between the maximum pixel intensity level of an image and its color channels is considered as the attenuation disparity of color channels. So, this leads to the statistical classification of underwater images into five major types: haze, bluish, greenish, greenish blue and bluish green images and two special types such as yellowish image and dark image subjectively. The pixel intensity of severely and moderately absorbed color channels are manipulated in accordance with the informative color channel by adding the value of concerned attenuation disparity globally. It increases the pixel intensity level of weak color channels and restores the channel information but does not improve the visual quality. The Gray World (GW) method or color equalization (CE) method is applied to remove the color imbalance and the contrast is improved by saturating the high level and low level pixel intensities. Finally, the conventional multiscale image fusion is employed to improve the image quality. The experimental results show that the proposed method outperforms qualitatively and quantitatively in terms of underwater color image quality evaluation (UCIQE) metrics and underwater image quality measures (UIQM) compared to the existing image enhancement methods.
- Open Access
- 26 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 32266-32277
Adjustment of Pixel Intensity for Removal of Color Casts in Underwater Images
This article introduces a new color correction method based on the adjustment of pixel intensity to remove color casts in underwater images. As the underwater images are captured in non-homogeneous with nonuniform illumination, the Gray-World Assumption (GWA) causes red artefacts. The method mainly aims to overcome the drawback of GWA for white balancing the underwater images. Before employing GWA, the underwater image is preprocessed by adding a color inferiority-reducing factor to the inferior channels to increase the pixel intensity level. The GWA is applied to preserve the information for substantially removing color casts. The image is adjusted by saturating the 1% top and bottom intensity levels pixels to remove the water haze. The color dominancy balanced image is finetuned by multi-scale image fusion technique with three derived inputs. The proposed method outperforms qualitatively and quantitatively based on the comparative analysis with four existing methods.
- 26 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 32266-32277
Open AccessAdjustment of Pixel Intensity for Removal of Color Casts in Underwater Images
This article introduces a new color correction method based on the adjustment of pixel intensity to remove color casts in underwater images. As the underwater images are captured in non-homogeneous with nonuniform illumination, the Gray-World Assumption (GWA) causes red artefacts. The method mainly aims to overcome the drawback of GWA for white balancing the underwater images. Before employing GWA, the underwater image is preprocessed by adding a color inferiority-reducing factor to the inferior channels to increase the pixel intensity level. The GWA is applied to preserve the information for substantially removing color casts. The image is adjusted by saturating the 1% top and bottom intensity levels pixels to remove the water haze. The color dominancy balanced image is finetuned by multi-scale image fusion technique with three derived inputs. The proposed method outperforms qualitatively and quantitatively based on the comparative analysis with four existing methods.
- 16 February 2024
- Published by Springer Nature in International Journal of Theoretical Physics
- Vol. 63 (2) , 1-26
Secured Quantum Image Communication Using New Two Dimensional Chaotic Map Based Encryption Methods
- 16 February 2024
- Published by Springer Nature in International Journal of Theoretical Physics
- Vol. 63 (2) , 1-26
Secured Quantum Image Communication Using New Two Dimensional Chaotic Map Based Encryption Methods