Background: Accurate classification of brain tumors in medical images is vital for effective diagnosis and treatment planning, which improves the patient’s survival rate. In this paper, we investigate the application of Convolutional Neural Networks (CNN) as a powerful tool for enhancing diagnostic accuracy using a Magnetic Resonance Imaging (MRI) dataset. Method: This study investigates the application of CNNs for brain tumor classification using a dataset of Magnetic Resonance Imaging (MRI) with a resolution of 200 × 200 × 1. The dataset is pre-processed and categorized into three types of tumors: Glioma, Meningioma, and Pituitary. The CNN models, including the Classic layer architecture and the ResNet50 architecture, are trained and evaluated using an 80:20 training-testing split. Results: The results reveal that both architectures accurately classify brain tumors. Classic layer architecture achieves an accuracy of 94.55%, while the ResNet50 architecture surpasses it with an accuracy of 99.88%. Compared to previous studies and 99.34%, our approach offers higher precision and reliability, demonstrating the effectiveness of ResNet50 in capturing complex features. Conclusions: The study concludes that CNNs, particularly the ResNet50 architecture, exhibit effectiveness in classifying brain tumors and hold significant potential in aiding medical professionals in accurate diagnosis and treatment planning. These advancements aim to further enhance the performance and practicality of CNN-based brain tumor classification systems, ultimately benefiting healthcare professionals and patients. For future research, exploring transfer learning techniques could be beneficial. By leveraging pre-trained models on large-scale datasets, researchers can utilize knowledge from other domains to improve brain tumor classification tasks, particularly in scenarios with limited annotated data.

A Good Performance OTP Encryption Image based on DCT-DWT Steganography

Klasifikasi Jeruk Nipis Terhadap Tingkat Kematangan Buah Berdasarkan Fitur Warna Menggunakan K-Nearest Neighbor

This paper proposed a hybrid method that had been solved both of these issues. We proposed a combination of Discrete Cosine Transform (DCT) and Haar Wavelet Transform (HWT) due to achieved imperceptibility and robustness comes from PSNR and NC value. Firstly, HWT applied in host image and then selected subband LL. LL subband will be divided into an 8×8 matrix in which each matrix was applied using DCT. Copyright embedded at the DC coefficient. In this paper we choose PSNR to measure the imperceptibility. Meanwhile, to measure the extracted copyright we used correlation coefficient. The results of this simulation method getting high PSNR value with an average of 43.2392 dB. Experiment results of extraction has tested using image watermarking attacks are JPEG compression, low pass filtering, mid filtering, unsharp filter, gaussian noise, cropingp, blurring, salt and pepper with the average value of the correlation coefficient of 0.8676. Our results proved that HWT-DCT has fulfilled the purpose of image watermarking, namely copyright protection.

Discrete Tchebichef Transform (DTT) is a transformation that has been widely applied in image processing, but still not widely applied in image watermarking. Based on the previous research results, DTT has advantages to reduce computational complexity, so that computation can be done faster and efficiently than the Discrete Cosine Transform (DCT). DTT also have similar properties with DCT, such as separability, symmetry, and orthogonality. DTT has also been shown to have higher energy compactness than the DCT, thus making DTT potential to be implemented in the image watermarking algorithm. In this paper, DTT will be introduced and applied in the techniques of image watermarking. Based on the results of this research, there is a very significant difference in the time required for computation, where DTT is proven to be very efficient to reduce algorithm complexity so that the time spent embedding and extracting copyright is about five times faster than DCT. Nevertheless, the watermarked image quality is also well preserved, as is the robust to various attacks.

Batik as one of Indonesia's cultural heritages has various types, motifs and colors. A batik may have almost the same motif with a different color or vice versa, therefore it requires a classification of batik motifs. In this study, a printed batik was used with various coastal batik motifs in Central Java. The algorithm for classification is selected Support Vector Machine (SVM) with feature extraction of the Gray Level Co-Occurrence Matrix (GLCM) and Local Binary Pattern (LBP). SVM has the advantage of grouping data with small amounts and short operation times. GLCM as an extractive feature for recognizing batik textures and LBP was chosen to do spot pattern recognition. In the experiment, we have used 160 images of batik motifs which are divided into two, namely 128 training data and 32 testing data. The accuracy results obtained from the SVM, GLCM and LBP algorithms produce 100% accuracy in polyniomial, linear and gaussian kernels with distances at GLCM 1, 3, and 5, where at a distance of 1 linear kernel is 78.1%, gaussian 93.7%. At a distance of 3 linear kernels 75%, gaussian 87.5% and at a distance of 5 linear kernels 84.3%, gaussian 87.5%. In the SVM and GLCM algorithms the resulting accuracy is at a distance of 1 with a polynomial kernel 96.8%, linear 68.7%, and gaussian 75%. At distance 3, the polynomial kernel is 100%, linear 71.8%, and gaussian 78.1%, while for distance 5, the polynomial kernel is 87.5%, linear 75%, and gaussian 81.2%.