Build a simple content analysis program

I have photographs of a corporate report for a company for 35 years - the photograph is entirely of text. There is a photograph of each page of each report; a total of 6,000 photographs organised into 35 different folders for the respective year of the corporate report. I want a piece of software where I can do the following: 1. Define the folder where the population of photographs are stored and from which a sample of photographs will be selected 2. User defines a sample size of X number of photos to be selected from the population randomly 3. I want the software to select a sample of photos following step (2.) and to produce a product key to identify that particular sample of photos so that specific sample can be used multiple times in the future. 4. Once the sample is selected, I want an option where the user enters in the "SAMPLEKEY" and the first photo in the sample appears on the screen; then users can click to go to the next photo in the sample. 4. Underneath each photo, as it appears on the screen, I need a categorization tree (which the software allows the user to input) with groups and subgroups which can go to multiple levels of subgroups if required. 5. The user is able to enter in a number next to each category and/or subcategory for each photo. 6. The software must be able to remember that particular sample from the SAMPLEKEY so multiple other users can click through the exact same sequence of photographs in the same order. 7. The software must be able to produce a report which presents the results of each users categorization results for each photo.