Piyush Jindal's Projects
My online portfolio and resume
A state of the art ensemble model for Handwritten Devanagari script OCR
Just to test the git setup in new environment
a web page (running on local host) with a button which when clicked would display a histogram with the following specifications. • The data for the histogram should be a random sample of 20 numbers taken without replacement from {1, 2, 3, . . . , 100} • The each block of histogram should be a shade of blue. • A block of histogram should turn orange when the mouse is placed on it. • When the mouse is rolled over a subset of adjacent blocks then they should appear in shades of orange.Further customization can b e done by asking visitors to the web p age for the following inputs and creating the histogram accordingly. • sample size • ◦ with replacement or ◦ without replacement • Normal with µ = and σ = • Uniform with minimum = and maximum = • Exponential with λ =
a web page (running on local host) with a button which when clicked would display a histogram with the following specifications. • The data for the histogram should be a random sample of 20 numbers taken without replacement from {1, 2, 3, . . . , 100} • The each block of histogram should be a shade of blue. • A block of histogram should turn orange when the mouse is placed on it. • When the mouse is rolled over a subset of adjacent blocks then they should appear in shades of orange.
I made this project under my R&D project and developed an algorithm to detect the size of foot using an image. I used Otsu's Global Thresholding for the segmentation and Circular Hough Transform(CHT) for reference object detection. I also developed an algorithm to find the start point and end point of the foot which may be present at any angle in the image to find the length of the foot. After finding the length, the model suggests the shoe size that will fit. I also proposed this algorithm for provision in online shopping to reduce the number of returns made due to the sizes . The whole program was written in MATLAB and a good accuracy was achieved.
word2vec implementation and training using gensim and applications.