Products(4) PST texture detection software

Product items

Image analysis / edge detection software developed by UCLA Prof. Bahram Jalali laboratory

Comparison with previous edge detection software

Feature of PST

Feature 1: textures and features are detected similarly without any influence of brightness

Feature 2: specimens are easily detected however illumination is insufficient

Feasibility test１: stable texture detection against different illumination level

Background:

As a texture value which is used in image based AI analysis and machine learning is based on pixel intensities, the value changes with illumination level. And a morphology value which is also used in AI analysis may change with boundary width.
Test result:

We prepared intentionally made 6 images with different illumination. However the texture values are varied with different illumination level without PST, the texture values are stabled with PST

Feasibility test 2: cell daughter analysis from dark images with 1/4 brightness

Background:

It is hard to make cell daughter analysis with photo toxic cells because cells have damages from illumination.
Test result:

By applying PST on time lapse images whose brightness level is darkened to 1/4 level, cell daughter　analysis of B cells is successively made. Collaborator:  Prof. Alexander Hoffmann, Signaling Systems Laboratory, UCLA

As a collaboration work with Prof. Jalali, Pinpoint Photonics, Inc. is searching business application of PST