{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Anomaly Detection in Camera Trap Images - Implementation\n",
    "This is an index file for the implementation part of my bachelor thesis 'Anomaly Detection in Camera Trap Images'."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Approach 1: Lapse Frame Differencing\n",
    "**Note:** Before this approach can be used on a new session, lapse maps have to be generated for this session using *scan_sessions.ipynb*.\n",
    "\n",
    " - *[approach1a_basic_frame_differencing.ipynb](approach1a_basic_frame_differencing.ipynb)*: Implementation.\n",
    " - *[approach1b_histograms.ipynb](approach1b_histograms.ipynb)*: Discarded similar approach using histogram distribution to compare Lapse and Motion images.\n",
    "\n",
    "## Approach 2: Median Frame Differencing\n",
    " - *[approach2_background_estimation.ipynb](approach2_background_estimation.ipynb)*: Implementation.\n",
    "\n",
    "## Approach 3: Bag of Visual Words\n",
    "### Notebooks\n",
    " - *[approach3_local_features.ipynb](approach3_local_features.ipynb)*: Visualizations and evaluation of single trainings.\n",
    " - *[approach3_boxplot.ipynb](approach3_boxplot.ipynb)*: Boxplot to compare multiple vocabularies generated using random prototypes.\n",
    "\n",
    "### Scripts\n",
    " - *[train_bow.py](train_bow.py)*: Training of BOW model\n",
    " - *[eval_bow.py](eval_bow.py)*: Evaluation of BOW model\n",
    "\n",
    "## Approach 4: Autoencoder\n",
    "### Notebooks\n",
    " - *[approach4_autoencoder.ipynb](approach4_autoencoder.ipynb)*: Visualizations and evaluation of single trainings.\n",
    " - *[approach4_boxplot.ipynb](approach4_boxplot.ipynb)*: Boxplot to compare different trainings.\n",
    "\n",
    "### Scripts\n",
    " - *[train_autoencoder.py](train_autoencoder.py)*: Training of autoencoder\n",
    " - *[eval_autoencoder.py](eval_autoencoder.py)*: Evaluation of trained autoencoder\n",
    "\n",
    "## Helpers\n",
    " - *[analyze_dataset.ipynb](analyze_dataset.ipynb)*: Dataset statistics, check for duplicates\n",
    " - *[analyze_labels.ipynb](analyze_labels.ipynb)*: Annotation statistics (number of normal/anomalous motion samples)\n",
    " - *[check_csv.ipynb](check_csv.ipynb)*: Loads annotations from *Kadaverbilder_leer.csv*\n",
    " - *[generate_lapseless_session.ipynb](generate_lapseless_session.ipynb)*: Generate a session with artificial lapse data from a lapseless session (e.g., Fox_03 -> GFox_03)\n",
    " - *[quick_label.py](quick_label.py)*: Minimal quick labeling script using OpenCV\n",
    " - *[read_csv_annotations.ipynb](read_csv_annotations.ipynb)*: Loads annotations from *observations.csv* and *media.csv*\n",
    " - *[resize_session.ipynb](resize_session.ipynb)*: Session preprocessing (crop and resize images)\n",
    " - *[scan_sessions.ipynb](scan_sessions.ipynb)*: Creates lapse maps (map between lapse images and their EXIF dates), statistics of inconsistencies in sessions\n",
    "\n",
    "## Early experiments\n",
    " - *[deprecated/experiments.ipynb](deprecated/experiments.ipynb)*: Early experiments with lapse images and frame differencing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Copyright © 2023 Felix Kleinsteuber and Computer Vision Group, Friedrich Schiller University Jena"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.10.4 ('pytorch-gpu')",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "name": "python",
   "version": "3.10.4"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "17cd5c528a3345b75540c61f907eece919c031d57a2ca1e5653325af249173c9"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}