Music Sheet Analysis Tool

---

Functions

• Crop Staves: Crop an input sheet music image into a set of individual lines of staves.
• Note Recognition: Identifies types of notes and signatures present in the line.
• Pitch Recognition: Classifies each detected note as a pitch (e.g., C4, G3) based on its position and clef.
• Midi Translation: Translates the extracted note, signature, and pitch information into MIDI format.

---

Code Packages

1. crop_staves

• Code for cropping staves into individual lines of music for downstream recognition and classification.

2. remove_staff_line

• Code for removing staff lines.

3. pitch_classification

• Code for determining the pitch of notes.

4. extract_symbol and recognition

• Code for note and signature classification.

---

Installing Dependencies

Install the required packages using the following command:

pip install -r requirements.txt

Note and Signature Recognition

Execute get_results.py after modifying the image_path in place in the code.

python get_results.py

Pitch Recognition

1. Run the Tool

Execute the main script with an input image:

python pitch_classification/main.py -i <path_to_image>

Replace <path_to_image> with the path to your input music sheet image (e.g., data/sample.png).

2. View Results

The outputs will be saved in the pitch_classification/output/ directory:

• Intermediate Results:
  • Processed images with lines or blobs removed.
  • Images with detected blobs and annotated staff numbers.
• Final Outputs:
  • Annotated image with detected notes and their pitches.
  • A text file (notes_pitch.txt) listing the detected notes and their pitches.

---

Key Parameters

• config.py:
  • VERBOSE: Enables detailed logging.
  • SAVING_IMAGES_STEPS: Saves intermediate image processing results.
  • THRESHOLD_MIN and THRESHOLD_MAX: Threshold values for binarizing the image.
  • LINES_DISTANCE_THRESHOLD: Minimum distance between staff lines for grouping.

---

File Outputs

• Annotated Images:
  • output/: Includes images for horizontal/vertical lines removed, detected blobs, and annotated notes with pitches.
• Text File:
  • output/notes_pitch.txt: Lists all detected notes and their pitches.

Info2MIDI Translation

Execute info2midi.py for an example of translation. Format the information in the format of the example, and call create_midi to execute the code on a custom file.