Enhancing image security: a lightweight crypto-steganographic approach for optimal quality

With the omnipresence of digital interaction, the need to secure sensitive information has increased tremendously. Hence, crypto-steganography has become an integral part of secret communication in various applications, ranging from image confidentiality in medical informatics to secured data transfer in sensor-based wireless networks. The challenge for researchers in IoT and embedded systems is to work under resource-constrained environments while preserving image quality and achieving computational efficiency. However, most of the current approaches consider cryptography and steganography as independent techniques, which often leads to higher computation overhead, reduced imperceptibility with increased payload, and low cryptographic security. Addressing these limitations, this work proposes a lightweight and unified crypto-steganographic approach whereby encryption is integrated into the embedding operation. A dual Piecewise Linear Chaotic Map (PWLCM) randomizes the reading order of payload and pixel embedding positions, while an in-place XOR transformation combines each payload bit with selected bits of pixels prior to embedding. The integration of Huffman compression effectively increases payload capacity, and multi-format support for lossless cover images (PNG, BMP, TIFF) and input file formats (. txt,. json,. csv) extends practical applicability. Experimental evaluation with both grayscale and color images shows very high imperceptibility, with peak PSNR values reaching 77. 42dB and SSIM of approximately 1. 0 for smaller payloads, while sustaining the PSNR above 51dB even for a full capacity of 166, 336 characters. Also, embedding and extraction remain less than 1. 6 seconds. Further, the robust statistical undetectability allows RS Analysis, Sample Pairs Analysis, and Chi-Square tests to be evaded even at high payload capacities. A key space of more than 1. 2483 2^199 ensures high cryptographic strength. All these observations together establish that the proposed approach is a lightweight embedding technique that can be used for big data, provides high cryptographic security and maintains image quality, hence is appropriate for low-resource situations.