def minimax(board, y, x, depth, isMax, alpha, beta):
if not isMax:
board[y][x] = 'b'
else:
board[y][x] = 'r'
if depth==0:
if isMax:
return evaluation(board, 'r', 'b', y, x)
else:
return evaluation(board, 'b', 'r', y, x)
if isMax:
best_score = -math.inf
moves = possible_moves(board)
for i, j in moves:
score = minimax(board, i, j, depth-1, not isMax, alpha, beta)
best_score = max(score, best_score)
alpha = max(alpha, best_score)
if alpha>=beta:
break
board[y][x] = ' '
return best_score
else:
best_score = math.inf
moves = possible_moves(board)
for i, j in moves:
score = minimax(board, i, j, depth-1, not isMax, alpha, beta)
best_score = min(-score, best_score)
beta = min(beta, best_score)
if alpha>=beta:
break
board[y][x] = ' '
return best_score
def itml_cmc(features, train_idxs, query_idxs, camId, gallery_idxs, labels):
N, m = features[train_idxs].shape
# features[train_idxs][:1000, ].shape
eigvals, eigvecs = calc_eig_pca_small(features[train_idxs].T, m, N)
m = 50
m_eigvecs = eigvecs[:, :m]
avg_face = compute_avg_face(features[train_idxs].T)
phi = features - avg_face
m_features = np.dot(phi, m_eigvecs)
itml = ITML_Supervised(verbose=True, num_constraints=5000, gamma=0.2)
X = m_features[train_idxs]
Y = labels[train_idxs]
X_itml = itml.fit_transform(X, Y)
M = itml.metric()
nn_idx_mat = evaluation(
knn,
features=m_features,
gallery_idxs=gallery_idxs,
query_idxs=query_idxs,
camId=camId,
labels=labels,
metric='mahalanobis',
metric_params={'VI': M}
)
return plot_CMC(nn_idx_mat, query_idxs, labels)
def lda_cmc(features, train_idxs, query_idxs, camId, gallery_idxs, labels):
N, m = features[train_idxs].shape
# features[train_idxs][:1000, ].shape
print("calculating PCA eigenvectors...")
eigvals, eigvecs = calc_eig_pca_small(features[train_idxs].T, m, N)
m = 1000
m_eigvecs = eigvecs[:, :m]
avg_face = compute_avg_face(features[train_idxs].T)
phi = features - avg_face
m_features = np.dot(phi, m_eigvecs)
print("calculating LDA eigenvectors...")
eigvals_lda, eigvecs_lda = calc_eig_lda(m_features[train_idxs].T, train_label=labels[train_idxs])
m_eigvecs_lda = eigvecs_lda[:,:len(set(labels[train_idxs]))-1].real
avg_face = compute_avg_face(m_features[train_idxs].T)
phi = m_features - avg_face
m_features_lda = np.dot(phi, m_eigvecs_lda)
print("evaluating LDA performance...")
nn_idx_mat = evaluation(
knn,
features=m_features_lda,
gallery_idxs=gallery_idxs,
query_idxs=query_idxs,
camId=camId,
labels=labels,
metric='euclidean',
metric_params=None
)
return plot_CMC(nn_idx_mat, query_idxs, labels)
def main():
print("importing data...")
data = loadmat('assets/cuhk03_new_protocol_config_labeled.mat')
with open('assets/feature_data.json') as f:
features = ujson.load(f)
print("data imported")
features = np.array(features)
train_idxs = data['train_idx'].flatten() - 1
query_idxs = data['query_idx'].flatten() - 1
camId = data['camId'].flatten()
gallery_idxs = data['gallery_idx'].flatten() - 1
labels = data['labels'].flatten()
N, m = features[train_idxs].shape
# features[train_idxs][:1000, ].shape
print("calculating PCA eigenvectors...")
eigvals, eigvecs = calc_eig_pca_small(features[train_idxs].T, m, N)
m = 1000
m_eigvecs = eigvecs[:, :m]
avg_face = compute_avg_face(features[train_idxs].T)
phi = features - avg_face
m_features = np.dot(phi, m_eigvecs)
print("calculating LDA eigenvectors...")
eigvals_lda, eigvecs_lda = calc_eig_lda(m_features[train_idxs].T,
train_label=labels[train_idxs])
m_eigvecs_lda = eigvecs_lda[:, :len(set(labels[train_idxs])) - 1].real
avg_face = compute_avg_face(m_features[train_idxs].T)
phi = m_features - avg_face
m_features_lda = np.dot(phi, m_eigvecs_lda)
print("evaluating LDA performance...")
nn_idx_mat = evaluation(knn,
features=m_features_lda,
gallery_idxs=gallery_idxs,
query_idxs=query_idxs,
camId=camId,
labels=labels,
metric='euclidean',
metric_params=None)
acc = get_all_rank_acc(nn_idx_mat, query_idxs, labels)
print("Accuracy:")
print(acc)
X = m_features_lda[train_idxs]
Y = labels[train_idxs]
plot_3d(X, Y)
test_set_idxs = np.append(gallery_idxs, query_idxs)
X_test = m_features_lda[test_set_idxs]
Y_test = labels[test_set_idxs]
n_cluster = np.unique(Y_test).size
nmi_kmean, acc_kmean = evaluation_k_means(X_test, n_cluster, Y_test)
print("PCA k-means accuracy (test set):")
print(acc_kmean)
def main():
print("importing data...")
data = loadmat('assets/cuhk03_new_protocol_config_labeled.mat')
with open('assets/feature_data.json') as f:
features = ujson.load(f)
print("data imported")
features = np.array(features)
train_idxs = data['train_idx'].flatten() - 1
query_idxs = data['query_idx'].flatten() - 1
camId = data['camId'].flatten()
gallery_idxs = data['gallery_idx'].flatten() - 1
labels = data['labels'].flatten()
print("evaluating baseline performance...")
nn_idx_mat = evaluation(knn,
features=features,
gallery_idxs=gallery_idxs,
query_idxs=query_idxs,
camId=camId,
labels=labels,
metric='euclidean',
metric_params=None)
acc = get_all_rank_acc(nn_idx_mat, query_idxs, labels)
print("Accuracy:")
print(acc)
X_ori = features[train_idxs]
Y_ori = labels[train_idxs]
plot_3d(X_ori, Y_ori)
n_cluster = np.unique(Y_ori).size
nmi_kmean, acc_kmean = evaluation_k_means(X_ori, n_cluster, Y_ori)
print("baseline k-means accuracy (train set):")
print(acc_kmean)
X_gallery = features[gallery_idxs]
Y_gallery = labels[gallery_idxs]
plot_3d(X_gallery, Y_gallery)
X_query = features[query_idxs]
Y_query = labels[query_idxs]
plot_3d(X_query, Y_query)
test_set_idxs = np.append(gallery_idxs, query_idxs)
X_test = features[test_set_idxs]
Y_test = labels[test_set_idxs]
n_cluster = np.unique(Y_test).size
nmi_kmean, acc_kmean = evaluation_k_means(X_test, n_cluster, Y_test)
print("baseline k-means accuracy (test set):")
print(acc_kmean)
def siamese_cmc(features, train_idxs, query_idxs, camId, gallery_idxs, labels):
siamese = Siamese()
siamese.load_model('3400')
siamese_out = siamese.test_model(input_1 = features)
nn_idx_mat = evaluation(
knn,
features=siamese_out,
gallery_idxs=gallery_idxs,
query_idxs=query_idxs,
camId=camId,
labels=labels,
metric='euclidean',
metric_params=None
)
return plot_CMC(nn_idx_mat, query_idxs, labels)
def pca_cmc(features, train_idxs, query_idxs, camId, gallery_idxs, labels):
N, m = features[train_idxs].shape
# features[train_idxs][:1000, ].shape
eigvals, eigvecs = calc_eig_pca_small(features[train_idxs].T, m, N)
m = 50
m_eigvecs = eigvecs[:, :m]
avg_face = compute_avg_face(features[train_idxs].T)
phi = features - avg_face
m_features = np.dot(phi, m_eigvecs)
print(m_features.shape)
print("evaluating PCA performance...")
nn_idx_mat = evaluation(
knn,
features=m_features,
gallery_idxs=gallery_idxs,
query_idxs=query_idxs,
camId=camId,
labels=labels,
metric='euclidean',
metric_params=None
)
return plot_CMC(nn_idx_mat, query_idxs, labels)
def main():
print("importing data...")
data = loadmat('assets/cuhk03_new_protocol_config_labeled.mat')
with open('assets/feature_data.json') as f:
features = ujson.load(f)
print("data imported")
features = np.array(features)
train_idxs = data['train_idx'].flatten() - 1
query_idxs = data['query_idx'].flatten() - 1
camId = data['camId'].flatten()
gallery_idxs = data['gallery_idx'].flatten() - 1
labels = data['labels'].flatten()
N, m = features[train_idxs].shape
# features[train_idxs][:1000, ].shape
eigvals, eigvecs = calc_eig_pca_small(features[train_idxs].T, m, N)
m = 50
m_eigvecs = eigvecs[:, :m]
avg_face = compute_avg_face(features[train_idxs].T)
phi = features - avg_face
m_features = np.dot(phi, m_eigvecs)
itml = ITML_Supervised(verbose=True, num_constraints=5000, gamma=0.1)
X = m_features[train_idxs]
Y = labels[train_idxs]
X_itml = itml.fit_transform(X, Y)
M = itml.metric()
plot_3d(X_itml, Y)
nn_idx_mat = evaluation(knn,
features=m_features,
gallery_idxs=gallery_idxs,
query_idxs=query_idxs,
camId=camId,
labels=labels,
metric='mahalanobis',
metric_params={'VI': M})
acc = get_all_rank_acc(nn_idx_mat, query_idxs, labels)
print("Accuracy:")
print(acc)
test_set_idxs = np.append(gallery_idxs, query_idxs)
features_ITML = itml.transform(m_features)
X_test = features_ITML[test_set_idxs]
Y_test = labels[test_set_idxs]
n_cluster = np.unique(Y_test).size
nmi_kmean, acc_kmean = evaluation_k_means(X_test, n_cluster, Y_test)
print("ITML k-means accuracy (test set):")
print(acc_kmean)
gamma = [i / 10 for i in range(1, 11)]
X_itmls = []
all_rank_acc_g = []
for g in gamma:
itml = ITML_Supervised(verbose=True, num_constraints=5000, gamma=0.2)
X = m_features[train_idxs]
X_itml = itml.fit_transform(X, Y)
X_itmls.append(X_itml)
M = itml.metric()
nn_idx_mat = evaluation(knn,
features=m_features,
gallery_idxs=gallery_idxs,
query_idxs=query_idxs,
camId=camId,
labels=labels,
metric='mahalanobis',
metric_params={'VI': M})
acc_g = get_all_rank_acc(nn_idx_mat, query_idxs, labels)
all_rank_acc_g.append(acc_g)
plt.plot(gamma, all_rank_acc_g)
plt.legend(('Rank 1', 'Rank 5', 'Rank10'))
plt.ylabel('Accuracy')
plt.xlabel('gamma')
print(all_rank_acc_g)
plt.show()