• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.11a
• /
• pp.112-115
• /
• 2022

The process of deep learning usually needs to deal with massive data which has greatly limited the development of deep learning technologies today. Convolutional Neural Network (CNN) structure is often used to solve image classification problems. However, a large number of images may be required in order to train an image in CNN, which is a heavy burden for existing computer systems to handle. If the image data can be compressed under the premise that the computer hardware system remains unchanged, it is possible to train more datasets in deep learning. However, image compression usually adopts the form of lossy compression, which will lose part of the image information. If the lost information is key information, it may affect learning performance. In this paper, we will analyze the effect of image compression on deep learning performance on CNN-based cat and dog classification. Through the experiment results, we conclude that the compression of images does not have a significant impact on the accuracy of deep learning.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.7 no.1
• /
• pp.142-148
• /
• 2003

We compress medical images in order to solve problems both of request of storage mediums and of a low network speed. In this paper, integration compression environment has been developed for unity of various compression methods. Various compression methods that are implemented by integration compression environment, RLC, Lossless JPEG, and JPEG, comply with the DICOM 3.0. A compression method using DWT is implemented at it. And a unit method of Lossless compression method and lossy compression method is designed to improve images quality and to progress compression ratio. Diverse medical images can be compressed by each compression method. And integration compression environment is operated together database so that information of medical images is administered.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.40 no.1
• /
• pp.45-51
• /
• 2003

JPEG200, a new international standard for still image compression based on wavelet and bit-plane coding techniques, is developed. In this paper, we design the DWT(Discrete Wavelet Transform) and quantizer for JPEG2000 CODEC. DWT handles both lossy and lossless compression using the same transform-based framework: The Daubechies 9/7 and 5/3 transforms, and quantizer is implemented as SQ(Scalar Quantization). The architecture of the proposed DWT and SQ are synthesized and verified using Xilinx FPGA technology. It operates up to 30MHz, and executes algorithms of wavelet transform and quantization for VGA 10 frame per second.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.13 no.3
• /
• pp.489-497
• /
• 2009

Most of existent blind watermarking methods in frequency domain use coefficient relationship to detect the watermark without original image. But the change in coefficient values occurred when the original image was attacked by lossy JPEG compression or noise addition. So robustness of watermark detection was weaken. In order to solve these problems, this paper presents a robust watermarking method, which enables multiple watermark insertion. Also, in order to reduce errors in the detected value of watermarks according to small changes in the coefficient relationship when detecting watermarks, it set the change guard value for variation of the coefficients. The experimental results show that the proposed method has good image quality and is robust to various attacks such as the JPEG lossy compression, noise addition, etc.

• Korean Journal of Computational Design and Engineering
• /
• v.13 no.1
• /
• pp.1-7
• /
• 2008

Data compression becomes increasingly an important issue for reducing data storage spaces as well as transmis-sion time in network environments. In 3D geometric models, the normal vectors of faces or meshes take a major portion of the data so that the compression of the vectors, which involves the trade off between the distortion of the images and compression ratios, plays a key role in reducing the size of the models. So, raising the compression ratio when the normal vector is compressed and minimizing the visual distortion of shape model's shading after compression are important. According to the recent papers, normal vector compression is useful to heighten com-pression ratio and to improve memory efficiency. But, the study about distortion of shading when the normal vector is compressed is rare relatively. In this paper, new normal vector compression method which is clustering normal vectors and assigning Representative Normal Vector (RNV) to each cluster and using the angular deviation from actual normal vector is proposed. And, using this new method, Visually Undistinguishable Lossy Compression (VULC) algorithm which distortion of shape model's shading by angular deviation of normal vector cannot be identified visually has been developed. And, being applied to the complicated shape models, this algorithm gave a good effectiveness.

• Journal of the Semiconductor & Display Technology
• /
• v.5 no.2 s.15
• /
• pp.19-25
• /
• 2006

This paper focuses on lossy medical image compression methods for medical images that operate on three-dimensional(3-D) irreversible integer wavelet transform. We offer an application of unbalanced tree structure algorithm to medical images, using a 3-D unbalanced wavelet decomposition and a 3-D unbalanced spatial dependence tree. The wavelet decomposition is accomplished with integer wavelet filters implemented with the lifting method. We have tested our encoder on volumetric medical images using different integer filters and coding unit sizes. The coding unit sizes of 16 slices save considerable dynamic memory(RAM) and coding delay from full sequence coding units used in previous works. If we allow the formation of trees of different lengths, then we can accomodate more transaxial scales than three. The encoder and decoder can then keep track of the length of the tree in which each pixel resides through the sequence of decompositions. Results show that, even with these small coding units, our algorithm with certain filters performs as well and better in lossy coding than previous coding systems using 3-D integer unbalanced wavelet transforms on volumetric medical images.

• Proceedings of the KOSOMBE Conference
• /
• v.1992 no.11
• /
• pp.29-33
• /
• 1992

We examine the compression performance of two irreversible (lossy) compression techniques, ADCT-VQ (Adaptive Discrete Cosine Trandform - Vector Quantization) and JPEG (Joint Photographic Experts group) which are basis of medical image information systems. Under the same compression ratio, MSE(Mean Square Error) is 0.578 lower in JPEG than in ADCT-VQ while SNR(Signal to Noise Ratio) is 1.236 dB higher in JPEG than in ADCT-VQ.

• Journal of Korea Multimedia Society
• /
• v.25 no.1
• /
• pp.29-38
• /
• 2022

Although image compression is one of the essential technologies to transmit image data on a variety of surveillance and mobile healthcare applications, it causes unnecessary compression artifacts such as blocking and ringing artifacts by the lossy compression in the limited network bandwidth. Recently, image restoration methods using convolutional neural network (CNN) show the significant improvement of image quality from the compressed images. In this paper, we propose Image Denoising Convolutional Neural Networks (IDCNN) to reduce the compression artifacts for the purpose of improving the performance of object classification. In order to evaluate the classification accuracy, we used the ImageNet test dataset consisting of 50,000 natural images and measured the classification performance in terms of Top-1 and Top-5 accuracy. Experimental results show that the proposed IDCNN can improve Top-1 and Top-5 accuracy as high as 2.46% and 2.42%, respectively.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10d
• /
• pp.454-456
• /
• 2002

압축 방법에는 크게 손실(lossy)압축과 무손실(lossless)압축으로 나눌 수 있다. 그 중 프랙탈 이미지 압축은 lossy 압축의 한가지 방법으로서 개별적인 화소들에 대한 자료를 저장하기보다는, 영상 생성을 위한 명령이나 방식을 저장하는 방법이다. 특히 이미지의 내에 자기유사성(self-similarity)과 중복성(Redundancy)을 이용하여 관련성을 발견하고 수학적인 공식으로 표현하려는 방식이다. 그러나 이미지를 Domain과 Range로 블록화 한 후 유사한 이미지를 찾아내는 데 걸리는 시간이 상당히 길다. 여기에서는 Domain과 Range의 외곽선의 기울기의 부호를 이용하여 블록을 16가지로 클래스화 하여서, 전체의 Domain 블록을 탐색하는 데 걸리는 시간을 줄이고자 하였다. 전체 탐색을 하는 경우보다 10배 이상의 속도향상을 보였고, 이미지에 따라서는 PSNR 값의 손실도 없음을 보였다.

• Proceedings of the Korean Society Of Semiconductor Equipment Technology
• /
• 2004.05a
• /
• pp.217-222
• /
• 2004

This paper focuses on lossy medical image compression methods for medical images that operate on two-dimensional(2D) integer wavelet transform. We offer an application of the Set Partitioning in Hierarchical Trees(SPIHT) algorithm to medical images, using a 2D wavelet decomposition and a 2D spatial dependence tree. The wavelet decomposition is accomplished with integer wavelet filters implemented with the lifting method, where careful scaling and truncations keep the integer precision small and the transform unitary. We have tested our encoder on medical images using different integer filters. Results show that our algorithm with certain filters performs as well and is sometimes better lossy coding using 2D integer wavelet transforms on medical images.