As the documentation says, collation affects only string comparisons, it doesn’t affect full-text matching, for this you need to adjust charset_table. What do you have in charset_table directive?
mysql> insert into idx_min(id,f) values(1, 'José María');
Query OK, 1 row affected (0.01 sec)
mysql> select * from idx_min where match('José');
+------+------+
| id | a |
+------+------+
| 1 | 0 |
+------+------+
1 row in set (0.00 sec)
mysql> select * from idx_min where match('Jose');
+------+------+
| id | a |
+------+------+
| 1 | 0 |
+------+------+
1 row in set (0.00 sec)
mysql> select * from idx_min where match('Maria');
+------+------+
| id | a |
+------+------+
| 1 | 0 |
+------+------+
1 row in set (0.00 sec)
Means numbers, latin, _ and russian characters are indexed (anything else is replaced with a whitespace). Also uppercases are mapped to lowercases (Apple and apple will match the same).
You can use the newly non_cjk alias, it’s a map with most UTF non-CJK characters (and includes mapping accents to latin form):