Meng-Hsiun Tsai

A multiple-level visual secret-sharing scheme without image size expansion

In traditional VSS schemes, the size of the share image is substantially expanded since each pixel of the secret image is mapped onto a block consisting of several pixels. In addition, the quality of the reconstructed secret image is normally degraded in contrast, especially for halftone images. This study proposes a VSS scheme that maps a block in a secret image onto one corresponding equal-sized block in each share image without image size expansion. Two types of techniques, including histogram width-equalization and histogram depth-equalization, are proposed to generate the corresponding share blocks containing multiple levels rather than two levels based on the density of black pixels on the blocks for a secret block. In the former technique, the gray-scale image histogram is obtained by uniformly splitting the range of the pixel gray levels in the secret image, while in the latter the buckets are created so that the area of each bucket is roughly constant by containing approximately the same number of pixels. The proposed schemes significantly improve the quality of the reconstructed secret image compared to several previous investigations.

A Decision Tree Based Classifier to Analyze Human Ovarian Cancer cDNA Microarray Datasets

Ovarian cancer is the deadliest gynaecological disease because of the high mortality rate and there is no any symptom in cancer early stage. It was often the terminal cancer period when patients were diagnosed with ovarian cancer and thus delays a good opportunity of treatment. The current common method for detecting ovarian cancer is blood testing for analyzing the tumor marker CA-125 of serum. However, specificity and sensitivity of CA-125 are insufficient for early detection. Therefore, it has become an urgent issue to look for an efficient method which precisely detects the tumor markers for ovarian cancer. This study aims to find the target genes of ovarian cancer by different algorithms of information science. Feature selection and decision tree were applied to analyze 9600 ovarian cancer-related genes. After screening the target genes, candidate genes will be analyzed by Ingenuity Pathway Analysis (IPA) software to create a genetic pathway model and to understand the interactive relationship in the different pathological stages of ovarian cancer. Finally, this research found 9 oncogenes associated with ovarian cancer and some genes had not been discovered in previous studies. This system will assist medical staffs in diagnosis and treatment at cancer early stage and improve the patient’s survival.

A protozoan parasite extraction scheme for digital microscopic images.

Pathogenic protozoan parasites can cause human to get many diseases, such as, amoebiasis, typhoid fever and cholera, etc. Different protozoan parasites vary greatly in their structural and biochemical properties. Digital images are extensively applied to medical fields for doctors and pathologists to analyze pathological sections and further diagnose diseases. The aim of this paper is to develop protozoan parasite extraction techniques to segment protozoan parasites from microscopic images. The proposed scheme has precise segmentation ability even if the image is with poor quality or complex background. Experimental results show that the proposed scheme can gain 96.64% average correct rate, and about 0.04, 0.45 and 0.06 of the average error rates: misclassification error (ME), region non-uniformity (RN) and relative foreground area error (RFAE), respectively.

Cymbidium mosaic potexvirus isolate-dependent host movement systems reveal two movement control determinants and the coat protein is the dominant.

Abstract Little is known about how plant viruses of a single species exhibit different movement behavior in different host species. Two Cymbidium mosaic potexvirus (CymMV) isolates, M1 and M2, were studied. Both can infect Phalaenopsis orchids, but only M1 can systemically infect Nicotiana benthamiana plants. Protoplast inoculation and whole-mount in situ hybridization revealed that both isolates can replicate in N. benthamiana; however, M2 was restricted to the initially infected cells. Genome shuffling between M1 and M2 revealed that two control modes are involved in CymMV host dependent movement. The M1 coat protein (CP) plays a dominant role in controlling CymMV movement between cells, because all chimeric CymMV viruses containing the M1 CP systemically infected N. benthamiana plants. Without the M1 CP, one chimeric virus containing the combination of the M1 triple gene block proteins (TGBps), the M2 5′ RNA (1–4333), and the M2 CP effectively moved in N. benthamiana plants. Further complementation analysis revealed that M1 TGBp1 and TGBp3 are co-required to complement the movement of the chimeric viruses in N. benthamiana. The amino acids within the CP, TGBp1 and TGBp3 which are required or important for CymMV M2 movement in N. benthamiana plants were mapped. The required amino acids within the CP map to the predicted RNA binding domain. RNA–protein binding assays revealed that M1 CP has higher RNA binding affinity than does M2 CP. Yeast two-hybrid assays to detect all possible interactions of M1 TGBps and CP, and only TGBp1 and CP self-interactions were observed.

A Database Watermarking Technique for Temper Detection

People pay much attention to the technology of data mining recently and more and more research institutions begin to buy the databases to analyze. If it doesn’t concern customer’s secrets the enterprises would also like to sell their data warehouse to do the research. Therefore, it becomes an important subject to prove the integrity of the database. This paper discusses about using the digital watermarking and the public authentication mechanism to strengthen the verification of integrity of the database. First, MD5 hash algorithm is used to fetch a database feature. Second, making XOR operation of database feature and digital watermarking gets a certification number. At last, using the secret key encrypts the certification number and makes public in the network with the database. Before using this database, user needs to use database owner’s public key to decrypt the ciphertext to get the certification number. Then making XOR operation of database feature fetched by MD5 algorithm and certification number gets a watermark. Finally, user can rely on the integrity of fetched watermark to understand whether the database is destroyed or not.

An Adaptable Threshold Decision Method

Otsu’s thresholding method (OTM) is one of the most commonly used thresholding methods. Unfortunately, the threshold obtained by OTM is biased in favor of the class, whose standard deviation or quantity of data is larger. Besides, one may adopt distinct thresholds in different applications for a same data set. Accordingly, this paper proposes an adaptable threshold decision method (ATDM) to provide the most appropriate thresholds for assorted applications. This paper also proposes a PSO (particle swarm optimization) based parameter detector (PBPD) to decide the fittest parameters which are used by ATDM. Image segmentation extracts the regions of interest from an image for follow-up analyses, and thresholding is one important technique for image segmentation. This paper will employ ATDM to detect the object contours in an image in order to investigate the performance of ATDM. The experiments show that ATDM can give impressive segmentation results.

Fragile Database Watermarking for Malicious Tamper Detection Using Support Vector Regression

This paper presents a digital watermarking technology for guaranteeing the database integrity. The proposed scheme based on the fragile watermarking technique, exploits trained support vector regression (SVR) predicting function to distribute the digital watermark over the particular numeric attributes to achieve embedding and detecting watermark by the same SVR predicting function. If the absolute value of the difference between predicted value and attribute value is more than the designed fixed value, like one, then the database content will be tampered with.

Blood Smear Image Based Malaria Parasite and Infected-Erythrocyte Detection and Segmentation

In this study, an automatic malaria parasite detector is proposed to perceive the malaria-infected erythrocytes in a blood smear image and to separate parasites from the infected erythrocytes. The detector hence can verify whether a patient is infected with malaria. It could more objectively and efficiently help a doctor in diagnosing malaria. The experimental results show that the proposed method can provide impressive performance in segmenting the malaria-infected erythrocytes and the parasites from a blood smear image taken under a microscope. This paper also presents a weighted Sobel operation to compute the image gradient. The experimental results demonstrates that the weighted Sobel operation can provide more clear-cut and thinner object contours in object segmentation.

A statistical and learning based oncogene detection and classification scheme using human cDNA expressions for ovarian carcinoma

In this paper, a human ovarian cDNA expression database is analyzed for detecting oncogenes and then selected oncogenes are used to identify pathological stages of ovarian carcinoma. This human ovarian cDNA expression database collects 41 patient samples which includes 13 samples of normal ovarian tumors (OVT), six samples of borderline of cancers (BOT), seven samples of ovarian cancer at stage I (OVCA-I) and 15 samples of ovarian cancer at stage III (OVCA-III). Each pathological sample contains a large number of genes (9600 genes). Hence oncogene analyzing and discovering is difficult. For this reason, a statistical testing method, t-test, is used to cull most of unconcerned genes in five different pathological stage classification cases. Then, these selected oncogenes are further used by artificial neural network (ANN) with five different classifications according to their gene expressions of pathological stages to set up a recognition system. This recognition system is used to show the efficiency of the proposed classification scheme. From the experimental results, the highest and lowest accuracy of five classification experiments is 100% and 89.47%. Moreover, this paper also proposed a novel t-test strategy to select more important oncogenes and increase lowest classification accuracy to 94.74%. The proposed scheme also can be used to develop a graphical user interface (GUI) bio-statistical or automatic diagnosis system for gene expression analysis to assist doctors and pathologists to analyze and diagnose ovarian cancer.

Oncogenes and Subtypes of Diffuse Large B-Cell Lymphoma Discoveries from Microarray Database

This paper presents an effective analysis scheme for Diffuse Large B-Cell Lymphoma (DLBCL) microarray datasets. Analysis of variable (ANOVA) is a well known statistics tools. It is useful to get the oncogenes to distinguish the normal and cancerous tissues. But, it can not further obtain the sub-types of cancerous tissues effectively. Hierarchical clustering is a well known analysis method for data mining. Therefore, it is also useful and fit to classify oncogenes to obtain some sub-types. ANOVA and hierarchical clustering both are employed to help us analyze B-cell Lymphoma datasets. In our analysis results, ANOVA can obtain 11 oncogenes of DLBCL from Stanford DLBCL microarray database successfully and accurately. Then, the 11 oncogenes are used for hierarchical clustering to identify the subtypes of cancerous tissues. In our hierarchical clustering analysis, we use 20 GC B-like DLBCL and 15 Activated B-like DLBCL actual samples used for analyzing. The analysis result shows that the hierarchical clustering can distinguish GC B-like DLBCL and Activated B-like DLBCL samples successfully.