Effloresce Cerebrum

Think of Effloresce Cerebrum as the API bridge designed in C++ between Effloresce Cordis and Effloresce Stomachum, its three major roles are to collect and process data for a first render for Effloresce Oculos, Morus and Nuntium, and collect and organize the data on Effloresce Stomachum, then provide a render package to Cordis.

We are using Caffe as the main deep learning framework, Caffe was developed by the Berkeley Vision and Learning Center (BVLC) that is written in C++ and designed for speed, modularity, and scalability. It provides a flexible architecture for designing and training deep neural networks, with support for convolutional neural networks (CNNs), recurrent neural networks (RNNs), and more.

Caffe supports data augmentation techniques to increase the diversity and size of the training dataset, which helps improve the generalization and robustness of the trained model. Data augmentation can be applied directly within Caffe using data augmentation layers or as part of the preprocessing pipeline before feeding the data into the network.

Next Steps:

Framework: Effloresce Cerebrum Framework Version: 0.01a Framework Update: April 2 2024 Framework: Effloresce Cerebrum Framework Version: 1.0b WORK: Dlib face detection improvment with object detection. Facial Expression analysis in Dlib, compress to simple Qizy3. Package dilivery without modifiers in data. GUI update for status in PHP, store status time data in SQL. Raw file compression when not in use. Bias reduction in backgrounds / seperate pillars from front object to background

Work Flow:

Pillar Data - Machine Learning

Pillars are the foundation of how we store data and bring data in to cerebrum for the render files, this process is started with data collection. Effloresce cerebrum is always learning via its own prompts with a program called brainstorm, this program creates prompts for cerebrum using internet data collection and a robot collector, it can also be manually prompted. (If Oculos is sent a GUI prompt it is sent to cerebrum with a top priority and brainstorm is run, first checking if it already has created pillars for a prompt)

A Oculos prompt will use Dlib frameworks of: Feature extraction, Object tracking, facial Expression, shape prediction, image segmentation, object detection, facial landmark detection, face detection and lesser Dlib vision frame works. It will pull the pillars for the prompt and match the pillars to datasets in SQL to create a render file, that will be sent to cordis AI for a second render

Here is a very simplified version of 7 pillars (when cerebrum pulled this in March 2024, it used 199 different pillars) of how pillars work for the prompt: "closeup portrait of a pretty 21 year old Caucasian white girl with freckles, long black hair"

Pillar 1 pulled will give cerebrum 1.9 million samples of what "closeup portrait" means and will have data sets for how many times each is used, and if this Pillar has been used in the past with other Pillers pulled in this render. With the pillars compiled and cordis computes and oculos renders if a change happens in the GUi prompt the changes are sent back to cerebrum and re-compiled and recorded for future reference and stores user data and statistics to suport the user to their preference.

Pillar Framework Examples:

Hair Color Pillar

Age Pillar