[Project Addition]: Butterfly Classification using DL - with Web App

Butterfly Species Classification using DL/Dataset/README.md

@@ -0,0 +1,106 @@
+# Dataset 
+The dataset contains individual folder for 100 species of butterfly, the species are : 
+1. ADONIS
+2. AFRICAN GIANT SWALLOWTAIL
+3. AMERICAN SNOOT
+4. AN 88
+5. APPOLLO
+6. ARCIGERA FLOWER MOTH
+7. ATALA
+8. ATLAS MOTH
+9. BANDED ORANGE HELICONIAN
+10. BANDED PEACOCK
+11. BANDED TIGER MOTH
+12. BECKERS WHITE
+13. BIRD CHERRY ERMINE MOTH
+14. BLACK HAIRSTREAK
+15. BLUE MORPHO
+16. BLUE SPOTTED CROW
+17. BROOKES BIRDWING
+18. BROWN ARGUS
+19. BROWN SIPROETA
+20. CABBAGE WHITE
+21. CAIRNS BIRDWING
+22. CHALK HILL BLUE
+23. CHECQUERED SKIPPER
+24. CHESTNUT
+25. CINNABAR MOTH
+26. CLEARWING MOTH
+27. CLEOPATRA
+28. CLODIUS PARNASSIAN
+29. CLOUDED SULPHUR
+30. COMET MOTH
+31. COMMON BANDED AWL
+32. COMMON WOOD-NYMPH
+33. COPPER TAIL
+34. CRECENT
+35. CRIMSON PATCH
+36. DANAID EGGFLY
+37. EASTERN COMA
+38. EASTERN DAPPLE WHITE
+39. EASTERN PINE ELFIN
+40. ELBOWED PIERROT
+41. EMPEROR GUM MOTH
+42. GARDEN TIGER MOTH
+43. GIANT LEOPARD MOTH
+44. GLITTERING SAPPHIRE
+45. GOLD BANDED
+46. GREAT EGGFLY
+47. GREAT JAY
+48. GREEN CELLED CATTLEHEART
+49. GREEN HAIRSTREAK
+50. GREY HAIRSTREAK
+51. HERCULES MOTH
+52. HUMMING BIRD HAWK MOTH
+53. INDRA SWALLOW
+54. IO MOTH
+55. Iphiclus sister
+56. JULIA
+57. LARGE MARBLE
+58. LUNA MOTH
+59. MADAGASCAN SUNSET MOTH
+60. MALACHITE
+61. MANGROVE SKIPPER
+62. MESTRA
+63. METALMARK
+64. MILBERTS TORTOISESHELL
+65. MONARCH
+66. MOURNING CLOAK
+67. OLEANDER HAWK MOTH
+68. ORANGE OAKLEAF
+69. ORANGE TIP
+70. ORCHARD SWALLOW
+71. PAINTED LADY
+72. PAPER KITE
+73. PEACOCK
+74. PINE WHITE
+75. PIPEVINE SWALLOW
+76. POLYPHEMUS MOTH
+77. POPINJAY
+78. PURPLE HAIRSTREAK
+79. PURPLISH COPPER
+80. QUESTION MARK
+81. RED ADMIRAL
+82. RED CRACKER
+83. RED POSTMAN
+84. RED SPOTTED PURPLE
+85. ROSY MAPLE MOTH
+86. SCARCE SWALLOW
+87. SILVER SPOT SKIPPER
+88. SIXSPOT BURNET MOTH
+89. SLEEPY ORANGE
+90. SOOTYWING
+91. SOUTHERN DOGFACE
+92. STRAITED QUEEN
+93. TROPICAL LEAFWING
+94. TWO BARRED FLASHER
+95. ULYSES
+96. VICEROY
+97. WHITE LINED SPHINX MOTH
+98. WOOD SATYR
+99. YELLOW SWALLOW TAIL
+100. ZEBRA LONG WING
+
+
+Link to dataset : https://www.kaggle.com/datasets/gpiosenka/butterfly-images40-species
+

Butterfly Species Classification using DL/Model/model.py

@@ -0,0 +1,63 @@
+import tensorflow as tf
+        from keras.preprocessing.image import ImageDataGenerator
+        train_datagen = ImageDataGenerator(rescale = 1./255,
+                                        shear_range = 0.2,
+                                        zoom_range = 0.2,
+                                        horizontal_flip = True)
+        training_set = train_datagen.flow_from_directory('dataset/train',
+                                                        target_size = (64, 64),
+                                                        batch_size = 30,
+                                                        class_mode = 'categorical')
+        test_datagen = ImageDataGenerator(rescale = 1./255)
+        test_set = test_datagen.flow_from_directory('dataset/test',
+                                                    target_size = (64, 64),
+                                                    batch_size = 30,
+                                                    class_mode = 'categorical')
+        # LeNet-5 architecture
+        lenet = tf.keras.models.Sequential()
+
+        # Layer 1: Convolutional layer with 6 filters, kernel size 5x5, and ReLU activation
+        lenet.add(tf.keras.layers.Conv2D(filters=6, kernel_size=5, activation='relu', input_shape=[64, 64, 3]))
+
+        # Layer 2: Average pooling layer with pool size 2x2 and strides 2
+        lenet.add(tf.keras.layers.AveragePooling2D(pool_size=2, strides=2))
+
+        # Layer 3: Convolutional layer with 16 filters, kernel size 5x5, and ReLU activation
+        lenet.add(tf.keras.layers.Conv2D(filters=16, kernel_size=5, activation='relu'))
+
+        # Layer 4: Average pooling layer with pool size 2x2 and strides 2
+        lenet.add(tf.keras.layers.AveragePooling2D(pool_size=2, strides=2))
+
+        # Layer 5: Flatten layer
+        lenet.add(tf.keras.layers.Flatten())
+
+        # Layer 6: Fully connected layer with 120 units and ReLU activation
+        lenet.add(tf.keras.layers.Dense(units=120, activation='relu'))
+
+        # Layer 7: Fully connected layer with 84 units and ReLU activation
+        lenet.add(tf.keras.layers.Dense(units=84, activation='relu'))
+
+        # Layer 8: Output layer with 46 units (assuming it's the number of classes) and softmax activation
+        lenet.add(tf.keras.layers.Dense(units=100, activation='softmax'))
+
+        # Compile the model with Adam optimizer and categorical crossentropy loss
+        lenet.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])
+
+        lenet.fit(x = training_set, validation_data = test_set, epochs = 15)
+
+        # Part 4 - Making a single prediction
+        print(training_set.class_indices)
+        import numpy as np
+        from keras.preprocessing import image
+        if uploaded_file is not None:
+            try:
+                class_indices = {'ADONIS': 0, 'AFRICAN GIANT SWALLOWTAIL': 1, 'AMERICAN SNOOT': 2, 'AN 88': 3, 'APPOLLO': 4, 'ARCIGERA FLOWER MOTH': 5, 'ATALA': 6, 'ATLAS MOTH': 7, 'BANDED ORANGE HELICONIAN': 8, 'BANDED PEACOCK': 9, 'BANDED TIGER MOTH': 10, 'BECKERS WHITE': 11, 'BIRD CHERRY ERMINE MOTH': 12, 'BLACK HAIRSTREAK': 13, 'BLUE MORPHO': 14, 'BLUE SPOTTED CROW': 15, 'BROOKES BIRDWING': 16, 'BROWN ARGUS': 17, 'BROWN SIPROETA': 18, 'CABBAGE WHITE': 19, 'CAIRNS BIRDWING': 20, 'CHALK HILL BLUE': 21, 'CHECQUERED SKIPPER': 22, 'CHESTNUT': 23, 'CINNABAR MOTH': 24, 'CLEARWING MOTH': 25, 'CLEOPATRA': 26, 'CLODIUS PARNASSIAN': 27, 'CLOUDED SULPHUR': 28, 'COMET MOTH': 29, 'COMMON BANDED AWL': 30, 'COMMON WOOD-NYMPH': 31, 'COPPER TAIL': 32, 'CRECENT': 33, 'CRIMSON PATCH': 34, 'DANAID EGGFLY': 35, 'EASTERN COMA': 36, 'EASTERN DAPPLE WHITE': 37, 'EASTERN PINE ELFIN': 38, 'ELBOWED PIERROT': 39, 'EMPEROR GUM MOTH': 40, 'GARDEN TIGER MOTH': 41, 'GIANT LEOPARD MOTH': 42, 'GLITTERING SAPPHIRE': 43, 'GOLD BANDED': 44, 'GREAT EGGFLY': 45, 'GREAT JAY': 46, 'GREEN CELLED CATTLEHEART': 47, 'GREEN HAIRSTREAK': 48, 'GREY HAIRSTREAK': 49, 'HERCULES MOTH': 50, 'HUMMING BIRD HAWK MOTH': 51, 'INDRA SWALLOW': 52, 'IO MOTH': 53, 'Iphiclus sister': 54, 'JULIA': 55, 'LARGE MARBLE': 56, 'LUNA MOTH': 57, 'MADAGASCAN SUNSET MOTH': 58, 'MALACHITE': 59, 'MANGROVE SKIPPER': 60, 'MESTRA': 61, 'METALMARK': 62, 'MILBERTS TORTOISESHELL': 63, 'MONARCH': 64, 'MOURNING CLOAK': 65, 'OLEANDER HAWK MOTH': 66, 'ORANGE OAKLEAF': 67, 'ORANGE TIP': 68, 'ORCHARD SWALLOW': 69, 'PAINTED LADY': 70, 'PAPER KITE': 71, 'PEACOCK': 72, 'PINE WHITE': 73, 'PIPEVINE SWALLOW': 74, 'POLYPHEMUS MOTH': 75, 'POPINJAY': 76, 'PURPLE HAIRSTREAK': 77, 'PURPLISH COPPER': 78, 'QUESTION MARK': 79, 'RED ADMIRAL': 80, 'RED CRACKER': 81, 'RED POSTMAN': 82, 'RED SPOTTED PURPLE': 83, 'ROSY MAPLE MOTH': 84, 'SCARCE SWALLOW': 85, 'SILVER SPOT SKIPPER': 86, 'SIXSPOT BURNET MOTH': 87, 'SLEEPY ORANGE': 88, 'SOOTYWING': 89, 'SOUTHERN DOGFACE': 90, 'STRAITED QUEEN': 91, 'TROPICAL LEAFWING': 92, 'TWO BARRED FLASHER': 93, 'ULYSES': 94, 'VICEROY': 95, 'WHITE LINED SPHINX MOTH': 96, 'WOOD SATYR': 97, 'YELLOW SWALLOW TAIL': 98, 'ZEBRA LONG WING': 99}
+                test_image = image.load_img(uploaded_file, target_size = (64, 64))
+                test_image = image.img_to_array(test_image)
+                test_image = np.expand_dims(test_image, axis = 0)
+                result = lenet.predict(test_image)
+                prediction = lenet.predict(test_image)
+                predicted_class_index = np.argmax(prediction)
+                predicted_class_name = [key for key, value in class_indices.items() if value == predicted_class_index][0]
+
+                print(predicted_class_index)

Butterfly Species Classification using DL/README.md

@@ -0,0 +1,40 @@
+# Hindi Letter Classification
+
+## Introduction
+
+Butterfly diversity is essential for ecosystem health and stability. However, manual classification of butterfly species can be time-consuming and prone to errors. An automated system using deep learning can expedite this process, providing a valuable tool for researchers, conservationists, and butterfly enthusiasts.
+
+![Alt text](https://i.pinimg.com/736x/b9/d6/83/b9d683dd495a0aaeb80d5b3ab7c156f0.jpg "Butterfly Species")
+
+## Web Application
+
+Butterfly species classification is a challenging task due to the diverse appearances and intricate patterns of butterflies. Accurate classification aids in biodiversity studies, conservation efforts, and ecological research. This project proposes a Convolutional Neural Network (CNN) approach utilizing the LeNet-5 architecture for automated butterfly species classification. The model demonstrates promising accuracy in distinguishing among 100 different butterfly species, facilitating researchers and enthusiasts in identifying and studying butterflies.
+
+## Libraries used
+
+1. Numpy
+2. Keras
+3. Tensorflow
+4. Streamlit
+
+# Dataset 
+
+Link : https://www.kaggle.com/datasets/gpiosenka/butterfly-images40-species
+
+# How to run locally
+
+Install the necessary libraries using pip then
+open the project folder and
+run following command : 
+
+```python
+
+streamlit run file_name.py
+
+```
+
+# Snapshots
+
+![Alt text](https://i.postimg.cc/mrPdbPyH/Screenshot-2024-05-15-175311.png "Snapshot 1")
+![Alt text](https://i.postimg.cc/q7Ht7Y8x/Screenshot-2024-05-15-175340.png "Snapshot 2")
+![Alt text](https://i.postimg.cc/CLZdsY74/Screenshot-2024-05-15-175356.png "Snapshot 3")

Butterfly Species Classification using DL/Web app/web_app.py

@@ -0,0 +1,95 @@
+import streamlit as st
+
+st.title("Butterfly Species classification 🦋")
+st.divider()
+st.image("https://i.postimg.cc/T1YP6Nr7/s-l1200.jpg")
+st.subheader("More about Butterflies")
+st.write('''
+Butterflies 🦋 are nature's delicate dancers, fluttering gracefully through sunlit meadows and vibrant gardens. With their kaleidoscope of colors and gentle wings, they bring joy and wonder wherever they go. Their metamorphosis from tiny caterpillars to exquisite winged beauties symbolizes transformation and growth 🌱. Watching them flit from flower to flower is a reminder to embrace change and find beauty in every stage of life. 🌸✨''')
+st.divider()
+st.subheader("Introduction")
+st.write('''
+The web app is created using streamlit framework. It contains a heading, small introduction and then a image uploader.
+          After the user uploads image, the image goes to backend and respected class is predicted by the CNN model and then
+          the uploaded image along with prediction is showed. We can play with prediciton time and accuracy by changing batch_size,
+          number of epochs and using a different CNN architecture.
+
+''')
+st.divider()
+uploaded_file = st.file_uploader("Enter image to Predict", type=['png', 'jpg'])
+submit = st.button("Submit")
+st.write("It may take 2-3 minutes to predict the image")
+if submit:
+    if uploaded_file is not None:
+
+        import tensorflow as tf
+        from keras.preprocessing.image import ImageDataGenerator
+        train_datagen = ImageDataGenerator(rescale = 1./255,
+                                        shear_range = 0.2,
+                                        zoom_range = 0.2,
+                                        horizontal_flip = True)
+        training_set = train_datagen.flow_from_directory('dataset/train',
+                                                        target_size = (64, 64),
+                                                        batch_size = 30,
+                                                        class_mode = 'categorical')
+        test_datagen = ImageDataGenerator(rescale = 1./255)
+        test_set = test_datagen.flow_from_directory('dataset/test',
+                                                    target_size = (64, 64),
+                                                    batch_size = 30,
+                                                    class_mode = 'categorical')
+        # LeNet-5 architecture
+        lenet = tf.keras.models.Sequential()
+
+        # Layer 1: Convolutional layer with 6 filters, kernel size 5x5, and ReLU activation
+        lenet.add(tf.keras.layers.Conv2D(filters=6, kernel_size=5, activation='relu', input_shape=[64, 64, 3]))
+
+        # Layer 2: Average pooling layer with pool size 2x2 and strides 2
+        lenet.add(tf.keras.layers.AveragePooling2D(pool_size=2, strides=2))
+
+        # Layer 3: Convolutional layer with 16 filters, kernel size 5x5, and ReLU activation
+        lenet.add(tf.keras.layers.Conv2D(filters=16, kernel_size=5, activation='relu'))
+
+        # Layer 4: Average pooling layer with pool size 2x2 and strides 2
+        lenet.add(tf.keras.layers.AveragePooling2D(pool_size=2, strides=2))
+
+        # Layer 5: Flatten layer
+        lenet.add(tf.keras.layers.Flatten())
+
+        # Layer 6: Fully connected layer with 120 units and ReLU activation
+        lenet.add(tf.keras.layers.Dense(units=120, activation='relu'))
+
+        # Layer 7: Fully connected layer with 84 units and ReLU activation
+        lenet.add(tf.keras.layers.Dense(units=84, activation='relu'))
+
+        # Layer 8: Output layer with 46 units (assuming it's the number of classes) and softmax activation
+        lenet.add(tf.keras.layers.Dense(units=100, activation='softmax'))
+
+        # Compile the model with Adam optimizer and categorical crossentropy loss
+        lenet.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])
+
+        lenet.fit(x = training_set, validation_data = test_set, epochs = 15)
+
+        # Part 4 - Making a single prediction
+        print(training_set.class_indices)
+        import numpy as np
+        from keras.preprocessing import image
+        if uploaded_file is not None:
+            try:
+                class_indices = {'ADONIS': 0, 'AFRICAN GIANT SWALLOWTAIL': 1, 'AMERICAN SNOOT': 2, 'AN 88': 3, 'APPOLLO': 4, 'ARCIGERA FLOWER MOTH': 5, 'ATALA': 6, 'ATLAS MOTH': 7, 'BANDED ORANGE HELICONIAN': 8, 'BANDED PEACOCK': 9, 'BANDED TIGER MOTH': 10, 'BECKERS WHITE': 11, 'BIRD CHERRY ERMINE MOTH': 12, 'BLACK HAIRSTREAK': 13, 'BLUE MORPHO': 14, 'BLUE SPOTTED CROW': 15, 'BROOKES BIRDWING': 16, 'BROWN ARGUS': 17, 'BROWN SIPROETA': 18, 'CABBAGE WHITE': 19, 'CAIRNS BIRDWING': 20, 'CHALK HILL BLUE': 21, 'CHECQUERED SKIPPER': 22, 'CHESTNUT': 23, 'CINNABAR MOTH': 24, 'CLEARWING MOTH': 25, 'CLEOPATRA': 26, 'CLODIUS PARNASSIAN': 27, 'CLOUDED SULPHUR': 28, 'COMET MOTH': 29, 'COMMON BANDED AWL': 30, 'COMMON WOOD-NYMPH': 31, 'COPPER TAIL': 32, 'CRECENT': 33, 'CRIMSON PATCH': 34, 'DANAID EGGFLY': 35, 'EASTERN COMA': 36, 'EASTERN DAPPLE WHITE': 37, 'EASTERN PINE ELFIN': 38, 'ELBOWED PIERROT': 39, 'EMPEROR GUM MOTH': 40, 'GARDEN TIGER MOTH': 41, 'GIANT LEOPARD MOTH': 42, 'GLITTERING SAPPHIRE': 43, 'GOLD BANDED': 44, 'GREAT EGGFLY': 45, 'GREAT JAY': 46, 'GREEN CELLED CATTLEHEART': 47, 'GREEN HAIRSTREAK': 48, 'GREY HAIRSTREAK': 49, 'HERCULES MOTH': 50, 'HUMMING BIRD HAWK MOTH': 51, 'INDRA SWALLOW': 52, 'IO MOTH': 53, 'Iphiclus sister': 54, 'JULIA': 55, 'LARGE MARBLE': 56, 'LUNA MOTH': 57, 'MADAGASCAN SUNSET MOTH': 58, 'MALACHITE': 59, 'MANGROVE SKIPPER': 60, 'MESTRA': 61, 'METALMARK': 62, 'MILBERTS TORTOISESHELL': 63, 'MONARCH': 64, 'MOURNING CLOAK': 65, 'OLEANDER HAWK MOTH': 66, 'ORANGE OAKLEAF': 67, 'ORANGE TIP': 68, 'ORCHARD SWALLOW': 69, 'PAINTED LADY': 70, 'PAPER KITE': 71, 'PEACOCK': 72, 'PINE WHITE': 73, 'PIPEVINE SWALLOW': 74, 'POLYPHEMUS MOTH': 75, 'POPINJAY': 76, 'PURPLE HAIRSTREAK': 77, 'PURPLISH COPPER': 78, 'QUESTION MARK': 79, 'RED ADMIRAL': 80, 'RED CRACKER': 81, 'RED POSTMAN': 82, 'RED SPOTTED PURPLE': 83, 'ROSY MAPLE MOTH': 84, 'SCARCE SWALLOW': 85, 'SILVER SPOT SKIPPER': 86, 'SIXSPOT BURNET MOTH': 87, 'SLEEPY ORANGE': 88, 'SOOTYWING': 89, 'SOUTHERN DOGFACE': 90, 'STRAITED QUEEN': 91, 'TROPICAL LEAFWING': 92, 'TWO BARRED FLASHER': 93, 'ULYSES': 94, 'VICEROY': 95, 'WHITE LINED SPHINX MOTH': 96, 'WOOD SATYR': 97, 'YELLOW SWALLOW TAIL': 98, 'ZEBRA LONG WING': 99}
+                test_image = image.load_img(uploaded_file, target_size = (64, 64))
+                test_image = image.img_to_array(test_image)
+                test_image = np.expand_dims(test_image, axis = 0)
+                result = lenet.predict(test_image)
+                prediction = lenet.predict(test_image)
+                predicted_class_index = np.argmax(prediction)
+                predicted_class_name = [key for key, value in class_indices.items() if value == predicted_class_index][0]
+                st.image(uploaded_file)
+                st.write(predicted_class_name)
+                print(predicted_class_index)
+
+
+            except:
+                st.write("There is error in file provided")
+
+        elif uploaded_file is None:
+            st.markdown(":red[Please enter a image]")

Butterfly Species Classification using DL/requirements.txt

@@ -0,0 +1,6 @@
+streamlit
+numpy
+pandas
+matplotlib
+tensorflow>=2.7.0
+keras>=2.6.0