For a suspected forgery that involves the falsification of a document or its contents, the investigator will primarily analyze the document’s paper and ink in order to establish the authenticity of the subject under investigation. As a non-destructive and contactless technique, Hyperspectral Imaging (HSI) is gaining popularity in the field of forensic document analysis. HSI returns more information compared to conventional three channel imaging systems due to the vast number of narrowband images recorded across the electromagnetic spectrum. As a result, HSI can provide better classification results. In this publication, we present results of an approach known as the t-Distributed Stochastic Neighbor Embedding (t-SNE) algorithm, which we have applied to HSI paper data analysis. Even though t-SNE has been widely accepted as a method for dimensionality reduction and visualization of high dimensional data, its usefulness has not yet been evaluated for the classification of paper data. In this research, we present a hyperspectral dataset of paper samples, and evaluate the clustering quality of the proposed method both visually and quantitatively. The t-SNE algorithm shows exceptional discrimination power when compared to traditional PCA with k-means clustering, in both visual and quantitative evaluations.

No abstract

Ink classification is an active topic in historic and forensic document analysis. In this work, we compared the ink classification capabilities of five commonly used and wellproven similarity measures for classification of hyperspectral imaging (HSI). They are Spectral Angle Mapper (SAM), Spectral Correlation Mapper (SCM), Euclidean Distance (ED), Spectral Information Divergence (SID) and Binary Encoding (BE). These techniques were well explored in different fields of HSI; however, they are not investigated in the field of ink classification. This study reveals the ink classification capabilities of these similarity measures. A combination of different types and colors of inks from different manufactures were used to create sample text. The SAM obtained higher accuracy compared to other methods and also identified that, inks that have nearly similar spectral signatures, cause a decline in accuracies due to misclassification between similar classes.

Firmness is one of the most important quality measures of strawberries, and is related to other aspects of the fruit, such as flavour, ripeness and internal characteristics. The most popular method for measuring firmness is puncturing with a penetrometer, which is destructive and time-consuming. In the present study, we make an attempt to predict the firmness of strawberries in a fast, non-destructive and non-contact way using hyperspectral imaging (HSI) and data analysis with various regression techniques. The primary goal of this research is to investigate and compare the firmness prediction capability of seven prominent regression techniques. We have performed HSI data acquisition of 150 strawberries and optimised seven regression models using the spectral information to predict strawberry firmness. These models are linear, ridge, lasso, k-neighbours, random forest, support vector and partial least square regression. The res ults show that HSI data with regression models has the potential to predict firmness in a rapid, non-destructive manner. Out of these seven regression models, the k-neighbours regression model outperformed all other methods with a standard error of prediction of 0.14, which is better than that of the state-of-the-art results.

Melanoma is the deadliest form of skin cancer and it is the most rapidly spreading cancer in the world. An earlier detection of this kind of cancer is curable, hence earlier detection of melanoma is pre-eminent. Because of this fact, a lot of research is being done in this area especially in automatic detection of melanoma. In this paper, we are proposing an automatic melanoma detection system which utilize a combination of deep and hand-crafted features. We analyzed the impact of using a simpler and standard hand-crafted feature, in place of complex usual hand-crafted features eg. shape, texture, diameter, or some custom features. We used a convolutional neural network (CNN) known as deep residual network (ResNet) to extract the deep features and utilized the scale invariant feature descriptor (SIFT) as the hand-crafted feature. The experiments revealed that, combining SIFT did not improve the accuracy of the system, however we obtained higher accuracy than state of the art methods with our deep only solution.