[Silva and Gonçalvez, 2019][1]

• Dataset
• Experiments
• Getting Started
• Some Results
• License
• Acknowledgement
• Reference

This repository contains some of experiments used for evaluating the regression with cnn (Convolutional Neural Networks) for counting better leafs. The dataset was available of the contents Computer Vision Problems in Plant Phenotyping.

As result one paper was produced and publicated in the WorkShop of Computer Vision (2019) where you can ready in the following link(da Silva and Gonçalvez (2019)).

Note: this paper was writer 80% in Portuguese, so is crucial you understand this language.

Bellow is a description about the experiments:

• Python script and Jupyter notebook;
• Was used a desktop computer with Intel(R) Xeon(R) CPU [email protected] GHz, 64 GB memory, and NVIDIA Titan V graphics card (5120 Compute Unified Device Architecture - CUDA cores and 12 GB graphics memory). The methods were implemented using Keras-Tensorflow on the Ubuntu 18.04 operating system.

• Sklearn 0.22.2.

  $ pip install -U scikit-learn

• TensorFlow

  $ pip install tensorflow

• Keras

  $ pip install keras

• Matplotlib

  $ python -m pip install -U matplotlib

• Os

  $ pip install os-sys

• Skimage

  $ pip install scikit-image

For run the experiment you need to access the respective architecture folder first, you can choose bellow the best way for run.

If you need run jupyter:

    jupyter-notebook name_experiment.ipynb 

You can run using python script also:

    python name_experiment.py

Bellow, follow two results that have been applied to the test set using the architecture Xception, and the metrics used was MAE (Mean Absolute error) and R2 (Coefficient of Determination):

Thanks for the teacher PhD Wesley Nunes Gonçalvez for contributing for this project.

• MIT license

[1]:da Silva, Neemias Bucéli, and Wesley Nunes Gonçalves. "Regression in Convolutional Neural Networks applied to Plant Leaf Counting." Anais do XV Workshop de Visão Computacional. SBC, 2019.

Sincerely Neemias B. Silva.